一_Decision_Three.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 概述\n",
    "## 决策树是如何工作的\n",
    "决策树（Decision Tree）是一种非参数的有监督学习方法，它能够从一系列有特征和标签的数据中总结出决策规\n",
    "则，并用树状图的结构来呈现这些规则，以解决分类和回归问题。决策树算法容易理解，适用各种数据，在解决各\n",
    "种问题时都有良好表现，尤其是以树模型为核心的各种集成算法，在各个行业和领域都有广泛的应用。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 构建决策数的两个阶段"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">1. 训练阶段  \n",
    "从给定的训练数据集DB，构造出一颗决策树  \n",
    "`class = DecisionTree(DB)`   \n",
    "2. 分类阶段  \n",
    "从根开始，按照决策树的分类属性逐层往下划分，知道叶节点，获得概念(决策、分类)  \n",
    "`y = DecisionTree(x)`\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 相关概念"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/concept.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 决策树算法"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 构建决策树\n",
    "原则上讲，任意一个数据集上的所有特征都可以被拿来分\n",
    "枝，特征上的任意节点又可以自由组合，所以一个数据集上可以发展出非常非常多棵决策树，其数量可达指数级。在\n",
    "这些树中，总有那么一棵树比其他的树分类效力都好，那样的树叫做”全局最优树“。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|**关键概念：全局最优、局部最优**|\n",
    "|  :------------------------ |  \n",
    "|全局最优：经过组合形成的，整体来说分类效果最好的模型|\n",
    "|局部最优：每一次分枝的时候都向着更好的分类效果分枝，但无法确认如此生成的树在全局上是否是最优的|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "要在这么多棵决策树中去一次性找到分类效果最佳的那一棵是不可能的，如果通过排列组合来进行筛选，计算量过于\n",
    "大而且低效，因此我们不会这样做。相对的，机器学习研究者们开发了一些有效的算法，能够在合理的时间内构造出\n",
    "具有一定准确率的**次最优**决策树。这些算法基本都执行”**贪心策略**“，即通过局部的最优来达到我们相信是最接近全局\n",
    "最优的结果。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|关键概念：贪心算法|\n",
    "| :---------|\n",
    "|通过实现局部最优来达到接近全局最优结果的算法，所有的树模型都是这样的算法。|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 算法的核心解决问题"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. 如何从数据表中找出最佳点和最佳分支\n",
    "2. 如何让决策树停止生长，防止过拟合\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 选择根节点(信息增益最大的当成根节点)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 决策树-熵\n",
    ">熵：表示数据内部混乱程度。熵越大越混乱。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "相关符合表示：  \n",
    "$H(X), H(Y)：表示当前事情X,Y发生的不确定性。$  \n",
    "$P(X)，P(Y)：表示当前事情发生的概率。$  \n",
    "$P(X)越大 --> H(X)越小。(可以这么理解，一件事情发生的概率越大说明它的不确定性越小。)$  \n",
    "$P(X)越小 --> H(X)越大。(可以这么理解，一件事情发生的概率越小说明它的不确定性越大。)$  \n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "假设：$A = \\{1,2,3,4,1,2,1\\}$  数字表示A里面的类别  \n",
    "$B = \\{1,1,1,1,1,2\\}$\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "由于A的类别很多，杂乱无章，我们可以说A的熵值很大，不纯度高。   \n",
    "由于B的类别很少(1或者2), B相对于A而言，我们可以说B的熵值很小，不纯度低。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|重要概念：不纯度|  \n",
    "| :----- |\n",
    "|决策树的每个叶子节点中都会包含一组数据，在这组数据中，如果有某一类标签占有较大的比例，我们就说叶子\n",
    "节点“纯”，分枝分得好。某一类标签占的比例越大，叶子就越纯，不纯度就越低，分枝就越好。\n",
    "如果没有哪一类标签的比例很大，各类标签都相对平均，则说叶子节点”不纯“，分枝不好，不纯度高|"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 公式"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$熵= - \\displaystyle\\sum_{i=1}^{n}P_iln(P_i)$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/shang.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "理解：当P比较小时，对应的lnP会比较大，所以能得到较大是熵，说明当前集合的混乱程度比较大。  \n",
    "当P比较大时，对应的lnP会比较小，所以得到的熵值比较小，说明当前集合的混乱程度比较小。  \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Gini(基尼系数)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$Gini系数 = Gini(p) = \\displaystyle \\sum_{k=1}^{K}p_k(1-p_k) = 1 - \\displaystyle \\sum_{k=1}^{K}p_k^2  \n",
    "\\text{(p为某一类别发生的概率)}$  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 构造决策树的基本想法"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">构造决策树的基本想法是随着树深度的增加，节点的熵**迅速下降**。熵下降的速度越快越好，这样我们就可以构造一颗**最矮**的决策树。(能用少量有用的信息解决事情越好)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 示例"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "根据前面的特征(outlook, temperature, humidty, windy)，判断是否出去打篮球(play)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/example1.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision1.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1.计算原始数据的熵值\n",
    "在没有给定任何天气信息时，根据历史数据，我们能知道一天打球的概率是9/14， 不打的概率是5/14。此时的熵为：  \n",
    "$-\\dfrac{9}{14}log_2\\dfrac{9}{14} - \\dfrac{5}{14}log_2\\dfrac{5}{14} = 0.940$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. 计算候选节点的熵值"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "下面我们计算当已知变量outlook的值时，信息熵为多少？"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision2.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对其加权求和"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision3.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 信息增益"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">信息增益表示：熵值发生的变化。  \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision4.png)\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ID3：：信息增益算法"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "信息增益的缺点：一些特征里面存在的属性很多，但是每个属性对应的的样本很少，这样会造成很大的信息增益。  \n",
    "示例： 我们为每一个样本增加一个ID(ID与结论无关，只是数据的一个标志)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision5.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "现在我们先考虑把ID当成根节点：  \n",
    "那么ID对应的信息熵为：  \n",
    "$\\dfrac{1}{14}\\times(-1\\times log_2(1)) + \\cdots + \\dfrac{1}{14}\\times(-1\\times log_2(1)) = 0$  \n",
    "熵值为0，信息增益最大，我们知道ID只是对样本的一个编号，不会对结论造成影响。但是，通过计算得到的结果会让我们错误地把ID当成根节点。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## C4.5(信息增益率)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "对应每个属性除以它自身的信息增益再求累加和"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 评价函数"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$C(T) = \\displaystyle \\sum_{t∈leaf}N_tH(t)$  （类似于损失函数，越小越好）"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 连续值的分界选择(如，年龄)\n",
    "![](images/decision6.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 剪枝操作"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "为什么要剪枝？  \n",
    "如果不剪枝的话，很容易导致过拟合。   \n",
    "如果决策树过于庞大，它包含的信息可能在训练集上很不错，但在测试集上很容易导致过拟合了，把一些噪音也涵盖进去了。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 预剪枝"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">预剪枝：在构建决策树过程中，能提前停止。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. 设置最大深度\n",
    "2. 设置最小节点\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 后剪枝\n",
    ">后剪枝：构建完决策树后，才开始剪枝操作"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. 修改损失函数\n",
    "![](images/decision7.png)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 随机森林(相当于构造多颗决策树)\n",
    "## 相关概念\n",
    "### 采样方式\n",
    "1. Bootstraping：有放回采样  \n",
    "2. Bagging：有放回采样n个样本建立分类器"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 随机\n",
    "### 双重随机性\n",
    "1. 数据选择的随机性\n",
    "2. 特征随机性"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision8.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# sklearn中的决策树\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T09:31:00.670913Z",
     "start_time": "2019-11-29T09:31:00.652378Z"
    }
   },
   "source": [
    "## 模块sklearn.tree\n",
    "|tree.DecisionTreeClassifier| 分类树|\n",
    "| -------------------------|------ |\n",
    "| tree.DecisionTreeRegressor| 回归树|  \n",
    "| tree.export_graphviz      |将生成的决策树导出为DOT格式，画图专用|\n",
    "| tree.ExtraTreeClassifier  |高随机版本的分类树|  \n",
    "| tree.ExtraTreeRegressor   | 高随机版本的回归树|\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## sklearn的基本建模流程\n",
    "1. 实例化，建立评估模型对象\n",
    "2. 通过模型接口训练模型\n",
    "3. 通过模型接口提取需要的信息  \n",
    "对应代码如下："
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T09:34:26.628679Z",
     "start_time": "2019-11-29T09:34:26.431579Z"
    }
   },
   "source": [
    "```python\n",
    "from sklearn import tree  # 导入模块\n",
    "\n",
    "clf = tree.DecisionTreeClassifier()  # 实例化，可以调参\n",
    "clf = clf.fit(x_train, y_train)  # 训练模型\n",
    "result = clf.score(x_test, y_test)  # 评估模型\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 利用DecisionTreeClassifier创建红酒数据集"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "```python\n",
    "DecisionTreeClassifier()\n",
    "\n",
    "Init signature:  \n",
    "tree.DecisionTreeClassifier(  \n",
    "    criterion='gini',  \n",
    "    splitter='best',  \n",
    "    max_depth=None,  \n",
    "    min_samples_split=2,  \n",
    "    min_samples_leaf=1,  \n",
    "    min_weight_fraction_leaf=0.0,  \n",
    "    max_features=None,  \n",
    "    random_state=None,  \n",
    "    max_leaf_nodes=None,  \n",
    "    min_impurity_decrease=0.0,  \n",
    "    min_impurity_split=None,  \n",
    "    class_weight=None,  \n",
    "    presort=False,  \n",
    ")\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 重要参数解读\n",
    "#### criterion(评估标准)\n",
    "为了要将表格转化为一棵树，决策树需要找出最佳节点和最佳的分枝方法，对分类树来说，衡量这个“最佳”的指标\n",
    "叫做“不纯度”。通常来说，**不纯度越低，决策树对训练集的拟合越好**。现在使用的决策树算法在分枝方法上的核心\n",
    "大多是围绕在对某个不纯度相关指标的最优化上。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision9.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">Criterion：用于决定不纯度的计算方法。sklearn提供了两种选择：  \n",
    "1. entropy：信息熵\n",
    "2. gini：基尼系数  \n",
    "  \n",
    "$$Entropy(t) = - \\displaystyle \\sum_{i=0}^{c-1}p(i|t)log_2p(i|t)$$  \n",
    "$$Gini(t) = 1 - \\displaystyle \\sum_{i = 0}^{c - 1}p(i|t)^2$$\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "其中t代表给定的节点，i代表标签的任意分类， $p(i|t)$代表标签分类i在节点t上所占的比例。注意，当使用信息熵\n",
    "时，sklearn实际计算的是基于信息熵的信息增益(Information Gain)，即父节点的信息熵和子节点的信息熵之差。\n",
    "比起基尼系数，**信息熵对不纯度更加敏感，对不纯度的惩罚最强**。但是**在实际使用中，信息熵和基尼系数的效果基\n",
    "本相同。**信息熵的计算比基尼系数缓慢一些，因为基尼系数的计算不涉及对数。另外，因为信息熵对不纯度更加敏\n",
    "感，所以信息熵作为指标时，决策树的生长会更加“精细”，因此对于高维数据或者噪音很多的数据，信息熵很容易\n",
    "过拟合，基尼系数在这种情况下效果往往比较好。当模型拟合程度不足的时候，即当模型在训练集和测试集上都表\n",
    "现不太好的时候，使用信息熵。当然，这些不是绝对的。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "|参数 |criterion|\n",
    "| ----|:--------|\n",
    "|如何影响模型? |确定不纯度的计算方法，帮忙找出最佳节点和最佳分枝，不纯度越低，决策树对训练集的拟合越好|\n",
    "|可能的输入有哪些？| 不填默认基尼系数，填写gini使用基尼系数，填写entropy使用信息增益|\n",
    "|怎样选取参数？|通常就使用基尼系数<br>数据维度很大，噪音很大时使用基尼系数<br>维度低，数据比较清晰的时候，信息熵和基尼系数没区别<br>当决策树的拟合程度不够的时候，使用信息熵<br>两个都试试，不好就换另外一个"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### random_state & splitter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "random_state用来设置分枝中的随机模式的参数，默认None，在高维度时随机性会表现更明显，低维度的数据\n",
    "（比如鸢尾花数据集），随机性几乎不会显现。输入任意整数，会一直长出同一棵树，让模型稳定下来。\n",
    "splitter也是用来控制决策树中的随机选项的，有两种输入值，输入”best\"，决策树在分枝时虽然随机，但是还是会\n",
    "优先选择更重要的特征进行分枝（重要性可以通过属性feature_importances_查看），输入“random\"，决策树在\n",
    "分枝时会更加随机，树会因为含有更多的不必要信息而更深更大，并因这些不必要信息而降低对训练集的拟合。这\n",
    "也是防止过拟合的一种方式。当你预测到你的模型会过拟合，用这两个参数来帮助你降低树建成之后过拟合的可能\n",
    "性。当然，树一旦建成，我们依然是使用剪枝参数来防止过拟合\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 剪枝参数\n",
    "在不加限制的情况下，一棵决策树会生长到衡量不纯度的指标最优，或者没有更多的特征可用为止。这样的决策树\n",
    "往往会过拟合，这就是说，**它会在训练集上表现很好，在测试集上却表现糟糕**。我们收集的样本数据不可能和整体\n",
    "的状况完全一致，因此当一棵决策树对训练数据有了过于优秀的解释性，它找出的规则必然包含了训练样本中的噪\n",
    "声，并使它对未知数据的拟合程度不足。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "为了让决策树有更好的泛化性，我们要对决策树进行剪枝。剪枝策略对决策树的影响巨大，正确的剪枝策略是优化\n",
    "决策树算法的核心。sklearn为我们提供了不同的剪枝策略："
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### max_depth"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">限制树的最大深度，超过设定深度的树枝全部剪掉  \n",
    "\n",
    "这是用得最广泛的剪枝参数，在高维度低样本量时非常有效。决策树多生长一层，对样本量的需求会增加一倍，所\n",
    "以限制树深度能够有效地限制过拟合。在集成算法中也非常实用。实际使用时，建议从=3开始尝试，看看拟合的效\n",
    "果再决定是否增加设定深度"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### min_samples_leaf\n",
    "min_samples_leaf限定，一个节点在分枝后的每个子节点都必须包含至少min_samples_leaf个训练样本，否则分\n",
    "枝就不会发生，或者，分枝会朝着满足每个子节点都包含min_samples_leaf个样本的方向去发生\n",
    "一般搭配max_depth使用，在回归树中有神奇的效果，可以让模型变得更加平滑。这个参数的数量设置得太小会引\n",
    "起过拟合，设置得太大就会阻止模型学习数据。一般来说，建议从=5开始使用。如果叶节点中含有的样本量变化很\n",
    "大，建议输入浮点数作为样本量的百分比来使用。同时，这个参数可以保证每个叶子的最小尺寸，可以在回归问题\n",
    "中避免低方差，过拟合的叶子节点出现。对于类别不多的分类问题，=1通常就是最佳选择"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### min_samples_split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "min_samples_split限定，一个节点必须要包含至少min_samples_split个训练样本，这个节点才允许被分枝，否则\n",
    "分枝就不会发生"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### max_features & min_impurity_decrease\n",
    "一般max_depth使用，用作树的”精修“\n",
    "max_features限制分枝时考虑的特征个数，超过限制个数的特征都会被舍弃。和max_depth异曲同工，\n",
    "max_features是用来限制高维度数据的过拟合的剪枝参数，但其方法比较暴力，是直接限制可以使用的特征数量\n",
    "而强行使决策树停下的参数，在不知道决策树中的各个特征的重要性的情况下，强行设定这个参数可能会导致模型\n",
    "学习不足。如果希望通过降维的方式防止过拟合，建议使用PCA，ICA或者特征选择模块中的降维算法。\n",
    "min_impurity_decrease限制信息增益的大小，信息增益小于设定数值的分枝不会发生。这是在0.19版本中更新的\n",
    "功能，在0.19版本之前时使用min_impurity_split"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![](images/decision10.png)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 实操\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 调包"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:39.815736Z",
     "start_time": "2019-12-06T05:06:39.051725Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn import tree\n",
    "from sklearn.datasets import load_wine\n",
    "from sklearn.model_selection import train_test_split  \n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 加载数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:42.275684Z",
     "start_time": "2019-12-06T05:06:42.260436Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1.423e+01, 1.710e+00, 2.430e+00, ..., 1.040e+00, 3.920e+00,\n",
       "        1.065e+03],\n",
       "       [1.320e+01, 1.780e+00, 2.140e+00, ..., 1.050e+00, 3.400e+00,\n",
       "        1.050e+03],\n",
       "       [1.316e+01, 2.360e+00, 2.670e+00, ..., 1.030e+00, 3.170e+00,\n",
       "        1.185e+03],\n",
       "       ...,\n",
       "       [1.327e+01, 4.280e+00, 2.260e+00, ..., 5.900e-01, 1.560e+00,\n",
       "        8.350e+02],\n",
       "       [1.317e+01, 2.590e+00, 2.370e+00, ..., 6.000e-01, 1.620e+00,\n",
       "        8.400e+02],\n",
       "       [1.413e+01, 4.100e+00, 2.740e+00, ..., 6.100e-01, 1.600e+00,\n",
       "        5.600e+02]])"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 加载数据\n",
    "wine = load_wine() \n",
    "# 显示数据\n",
    "wine.data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 显示数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:45.289123Z",
     "start_time": "2019-12-06T05:06:45.279033Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,\n",
       "       0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,\n",
       "       1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2,\n",
       "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
       "       2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,\n",
       "       2, 2])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 显示数据的结果/分类\n",
    "wine.target"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:45.959051Z",
     "start_time": "2019-12-06T05:06:45.928883Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>alcohol</th>\n",
       "      <th>malic_acid</th>\n",
       "      <th>ash</th>\n",
       "      <th>alcalinity_of_ash</th>\n",
       "      <th>magnesium</th>\n",
       "      <th>total_phenols</th>\n",
       "      <th>flavanoids</th>\n",
       "      <th>nonflavanoid_phenols</th>\n",
       "      <th>proanthocyanins</th>\n",
       "      <th>color_intensity</th>\n",
       "      <th>hue</th>\n",
       "      <th>od280/od315_of_diluted_wines</th>\n",
       "      <th>proline</th>\n",
       "      <th>class</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>14.23</td>\n",
       "      <td>1.71</td>\n",
       "      <td>2.43</td>\n",
       "      <td>15.6</td>\n",
       "      <td>127.0</td>\n",
       "      <td>2.80</td>\n",
       "      <td>3.06</td>\n",
       "      <td>0.28</td>\n",
       "      <td>2.29</td>\n",
       "      <td>5.64</td>\n",
       "      <td>1.04</td>\n",
       "      <td>3.92</td>\n",
       "      <td>1065.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>13.20</td>\n",
       "      <td>1.78</td>\n",
       "      <td>2.14</td>\n",
       "      <td>11.2</td>\n",
       "      <td>100.0</td>\n",
       "      <td>2.65</td>\n",
       "      <td>2.76</td>\n",
       "      <td>0.26</td>\n",
       "      <td>1.28</td>\n",
       "      <td>4.38</td>\n",
       "      <td>1.05</td>\n",
       "      <td>3.40</td>\n",
       "      <td>1050.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>13.16</td>\n",
       "      <td>2.36</td>\n",
       "      <td>2.67</td>\n",
       "      <td>18.6</td>\n",
       "      <td>101.0</td>\n",
       "      <td>2.80</td>\n",
       "      <td>3.24</td>\n",
       "      <td>0.30</td>\n",
       "      <td>2.81</td>\n",
       "      <td>5.68</td>\n",
       "      <td>1.03</td>\n",
       "      <td>3.17</td>\n",
       "      <td>1185.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14.37</td>\n",
       "      <td>1.95</td>\n",
       "      <td>2.50</td>\n",
       "      <td>16.8</td>\n",
       "      <td>113.0</td>\n",
       "      <td>3.85</td>\n",
       "      <td>3.49</td>\n",
       "      <td>0.24</td>\n",
       "      <td>2.18</td>\n",
       "      <td>7.80</td>\n",
       "      <td>0.86</td>\n",
       "      <td>3.45</td>\n",
       "      <td>1480.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>13.24</td>\n",
       "      <td>2.59</td>\n",
       "      <td>2.87</td>\n",
       "      <td>21.0</td>\n",
       "      <td>118.0</td>\n",
       "      <td>2.80</td>\n",
       "      <td>2.69</td>\n",
       "      <td>0.39</td>\n",
       "      <td>1.82</td>\n",
       "      <td>4.32</td>\n",
       "      <td>1.04</td>\n",
       "      <td>2.93</td>\n",
       "      <td>735.0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   alcohol  malic_acid   ash  alcalinity_of_ash  magnesium  total_phenols  \\\n",
       "0    14.23        1.71  2.43               15.6      127.0           2.80   \n",
       "1    13.20        1.78  2.14               11.2      100.0           2.65   \n",
       "2    13.16        2.36  2.67               18.6      101.0           2.80   \n",
       "3    14.37        1.95  2.50               16.8      113.0           3.85   \n",
       "4    13.24        2.59  2.87               21.0      118.0           2.80   \n",
       "\n",
       "   flavanoids  nonflavanoid_phenols  proanthocyanins  color_intensity   hue  \\\n",
       "0        3.06                  0.28             2.29             5.64  1.04   \n",
       "1        2.76                  0.26             1.28             4.38  1.05   \n",
       "2        3.24                  0.30             2.81             5.68  1.03   \n",
       "3        3.49                  0.24             2.18             7.80  0.86   \n",
       "4        2.69                  0.39             1.82             4.32  1.04   \n",
       "\n",
       "   od280/od315_of_diluted_wines  proline  class  \n",
       "0                          3.92   1065.0      0  \n",
       "1                          3.40   1050.0      0  \n",
       "2                          3.17   1185.0      0  \n",
       "3                          3.45   1480.0      0  \n",
       "4                          2.93    735.0      0  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 数据处理，合并数据\n",
    "wine_df = pd.concat([pd.DataFrame(wine.data, columns=wine.feature_names),\n",
    "                    pd.DataFrame(wine.target, columns=['class'])], \n",
    "                    axis=1)\n",
    "wine_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:46.716396Z",
     "start_time": "2019-12-06T05:06:46.712214Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['alcohol',\n",
       " 'malic_acid',\n",
       " 'ash',\n",
       " 'alcalinity_of_ash',\n",
       " 'magnesium',\n",
       " 'total_phenols',\n",
       " 'flavanoids',\n",
       " 'nonflavanoid_phenols',\n",
       " 'proanthocyanins',\n",
       " 'color_intensity',\n",
       " 'hue',\n",
       " 'od280/od315_of_diluted_wines',\n",
       " 'proline']"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 显示特征名\n",
    "wine.feature_names  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:47.423885Z",
     "start_time": "2019-12-06T05:06:47.416656Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['class_0', 'class_1', 'class_2'], dtype='<U7')"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 显示分类\n",
    "wine.target_names"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 分训练集和测试集\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:06:48.909056Z",
     "start_time": "2019-12-06T05:06:48.894686Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(124, 13)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 一定要严格按照格式写，不然可能导数训练和测试乱了\n",
    "# test_size:表示测试集占总的数据的比例\n",
    "x_train, x_test, y_train, y_test = train_test_split(wine.data, \n",
    "                                                    wine.target,\n",
    "                                                    test_size=0.3\n",
    "                                                    )\n",
    "x_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 建立模型"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:00.087035Z",
     "start_time": "2019-12-06T05:07:00.072744Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9629629629629629"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = tree.DecisionTreeClassifier(criterion='entropy')  # 修改标准为信息增益\n",
    "clf = clf.fit(x_train, y_train)\n",
    "# 返回测试的准确度\n",
    "score = clf.score(x_test, y_test)\n",
    "\n",
    "score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T10:10:56.938196Z",
     "start_time": "2019-11-29T10:10:56.914412Z"
    }
   },
   "source": [
    "#### 画出一棵树"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:02.527184Z",
     "start_time": "2019-12-06T05:07:02.517956Z"
    }
   },
   "outputs": [],
   "source": [
    "import graphviz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:04.080194Z",
     "start_time": "2019-12-06T05:07:03.997472Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.40.1 (20161225.0304)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"667pt\" height=\"552pt\"\n",
       " viewBox=\"0.00 0.00 666.50 552.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 548)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"#ffffff\" stroke=\"transparent\" points=\"-4,4 -4,-548 662.5,-548 662.5,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\">\n",
       "<title>0</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.098039\" stroke=\"#000000\" d=\"M312,-544C312,-544 204,-544 204,-544 198,-544 192,-538 192,-532 192,-532 192,-473 192,-473 192,-467 198,-461 204,-461 204,-461 312,-461 312,-461 318,-461 324,-467 324,-473 324,-473 324,-532 324,-532 324,-538 318,-544 312,-544\"/>\n",
       "<text text-anchor=\"middle\" x=\"258\" y=\"-528.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">类黄酮 &lt;= 1.425</text>\n",
       "<text text-anchor=\"middle\" x=\"258\" y=\"-513.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 1.562</text>\n",
       "<text text-anchor=\"middle\" x=\"258\" y=\"-498.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 124</text>\n",
       "<text text-anchor=\"middle\" x=\"258\" y=\"-483.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 50, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"258\" y=\"-468.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\">\n",
       "<title>1</title>\n",
       "<path fill=\"#8139e5\" fill-opacity=\"0.780392\" stroke=\"#000000\" d=\"M238.5,-425C238.5,-425 131.5,-425 131.5,-425 125.5,-425 119.5,-419 119.5,-413 119.5,-413 119.5,-354 119.5,-354 119.5,-348 125.5,-342 131.5,-342 131.5,-342 238.5,-342 238.5,-342 244.5,-342 250.5,-348 250.5,-354 250.5,-354 250.5,-413 250.5,-413 250.5,-419 244.5,-425 238.5,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"185\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 3.725</text>\n",
       "<text text-anchor=\"middle\" x=\"185\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.679</text>\n",
       "<text text-anchor=\"middle\" x=\"185\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 39</text>\n",
       "<text text-anchor=\"middle\" x=\"185\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 7, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"185\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 贝尔摩德</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\">\n",
       "<title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M232.4681,-460.8796C227.109,-452.1434 221.4021,-442.8404 215.8718,-433.8253\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"218.8533,-431.9919 210.6408,-425.2981 212.8865,-435.6522 218.8533,-431.9919\"/>\n",
       "<text text-anchor=\"middle\" x=\"204.8161\" y=\"-445.9101\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">True</text>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\">\n",
       "<title>4</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.023529\" stroke=\"#000000\" d=\"M381.5,-425C381.5,-425 280.5,-425 280.5,-425 274.5,-425 268.5,-419 268.5,-413 268.5,-413 268.5,-354 268.5,-354 268.5,-348 274.5,-342 280.5,-342 280.5,-342 381.5,-342 381.5,-342 387.5,-342 393.5,-348 393.5,-354 393.5,-354 393.5,-413 393.5,-413 393.5,-419 387.5,-425 381.5,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">脯氨酸 &lt;= 724.5</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 1.0</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 85</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 43, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\">\n",
       "<title>0&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M283.5319,-460.8796C288.891,-452.1434 294.5979,-442.8404 300.1282,-433.8253\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"303.1135,-435.6522 305.3592,-425.2981 297.1467,-431.9919 303.1135,-435.6522\"/>\n",
       "<text text-anchor=\"middle\" x=\"311.1839\" y=\"-445.9101\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\">\n",
       "<title>2</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M98,-298.5C98,-298.5 12,-298.5 12,-298.5 6,-298.5 0,-292.5 0,-286.5 0,-286.5 0,-242.5 0,-242.5 0,-236.5 6,-230.5 12,-230.5 12,-230.5 98,-230.5 98,-230.5 104,-230.5 110,-236.5 110,-242.5 110,-242.5 110,-286.5 110,-286.5 110,-292.5 104,-298.5 98,-298.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 7</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 7, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\">\n",
       "<title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M139.5323,-341.8796C126.6789,-330.1138 112.7021,-317.3197 99.8834,-305.5855\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"102.2264,-302.9854 92.4869,-298.8149 97.4999,-308.1488 102.2264,-302.9854\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\">\n",
       "<title>3</title>\n",
       "<path fill=\"#8139e5\" stroke=\"#000000\" d=\"M236,-298.5C236,-298.5 140,-298.5 140,-298.5 134,-298.5 128,-292.5 128,-286.5 128,-286.5 128,-242.5 128,-242.5 128,-236.5 134,-230.5 140,-230.5 140,-230.5 236,-230.5 236,-230.5 242,-230.5 248,-236.5 248,-242.5 248,-242.5 248,-286.5 248,-286.5 248,-292.5 242,-298.5 236,-298.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"188\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"188\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 32</text>\n",
       "<text text-anchor=\"middle\" x=\"188\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 0, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"188\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 贝尔摩德</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\">\n",
       "<title>1&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M186.0493,-341.8796C186.3182,-331.2134 186.6084,-319.7021 186.8806,-308.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"190.3817,-308.9 187.1349,-298.8149 183.3839,-308.7235 190.3817,-308.9\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\">\n",
       "<title>5</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.976471\" stroke=\"#000000\" d=\"M375.5,-306C375.5,-306 282.5,-306 282.5,-306 276.5,-306 270.5,-300 270.5,-294 270.5,-294 270.5,-235 270.5,-235 270.5,-229 276.5,-223 282.5,-223 282.5,-223 375.5,-223 375.5,-223 381.5,-223 387.5,-229 387.5,-235 387.5,-235 387.5,-294 387.5,-294 387.5,-300 381.5,-306 375.5,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"329\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">酒精 &lt;= 13.175</text>\n",
       "<text text-anchor=\"middle\" x=\"329\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.165</text>\n",
       "<text text-anchor=\"middle\" x=\"329\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 41</text>\n",
       "<text text-anchor=\"middle\" x=\"329\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [1, 40, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"329\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\">\n",
       "<title>4&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M330.3005,-341.8796C330.1628,-333.6838 330.0166,-324.9891 329.874,-316.5013\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"333.3701,-316.2378 329.7025,-306.2981 326.3711,-316.3555 333.3701,-316.2378\"/>\n",
       "</g>\n",
       "<!-- 10 -->\n",
       "<g id=\"node11\" class=\"node\">\n",
       "<title>10</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.925490\" stroke=\"#000000\" d=\"M524.5,-306C524.5,-306 417.5,-306 417.5,-306 411.5,-306 405.5,-300 405.5,-294 405.5,-294 405.5,-235 405.5,-235 405.5,-229 411.5,-223 417.5,-223 417.5,-223 524.5,-223 524.5,-223 530.5,-223 536.5,-229 536.5,-235 536.5,-235 536.5,-294 536.5,-294 536.5,-300 530.5,-306 524.5,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"471\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 3.435</text>\n",
       "<text text-anchor=\"middle\" x=\"471\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.359</text>\n",
       "<text text-anchor=\"middle\" x=\"471\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 44</text>\n",
       "<text text-anchor=\"middle\" x=\"471\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [41, 3, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"471\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 琴酒</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;10 -->\n",
       "<g id=\"edge10\" class=\"edge\">\n",
       "<title>4&#45;&gt;10</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M379.9652,-341.8796C390.9847,-332.513 402.7709,-322.4948 414.0842,-312.8784\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"416.4732,-315.4414 421.8258,-306.2981 411.9396,-310.1078 416.4732,-315.4414\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\">\n",
       "<title>6</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M242.5,-179.5C242.5,-179.5 149.5,-179.5 149.5,-179.5 143.5,-179.5 137.5,-173.5 137.5,-167.5 137.5,-167.5 137.5,-123.5 137.5,-123.5 137.5,-117.5 143.5,-111.5 149.5,-111.5 149.5,-111.5 242.5,-111.5 242.5,-111.5 248.5,-111.5 254.5,-117.5 254.5,-123.5 254.5,-123.5 254.5,-167.5 254.5,-167.5 254.5,-173.5 248.5,-179.5 242.5,-179.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"196\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"196\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 38</text>\n",
       "<text text-anchor=\"middle\" x=\"196\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 38, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"196\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\">\n",
       "<title>5&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M282.4831,-222.8796C269.333,-211.1138 255.0337,-198.3197 241.9191,-186.5855\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"244.1382,-183.8746 234.352,-179.8149 239.4706,-189.0913 244.1382,-183.8746\"/>\n",
       "</g>\n",
       "<!-- 7 -->\n",
       "<g id=\"node8\" class=\"node\">\n",
       "<title>7</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.498039\" stroke=\"#000000\" d=\"M377.5,-187C377.5,-187 284.5,-187 284.5,-187 278.5,-187 272.5,-181 272.5,-175 272.5,-175 272.5,-116 272.5,-116 272.5,-110 278.5,-104 284.5,-104 284.5,-104 377.5,-104 377.5,-104 383.5,-104 389.5,-110 389.5,-116 389.5,-116 389.5,-175 389.5,-175 389.5,-181 383.5,-187 377.5,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">脯氨酸 &lt;= 655.0</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.918</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 3</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [1, 2, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"331\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;7 -->\n",
       "<g id=\"edge7\" class=\"edge\">\n",
       "<title>5&#45;&gt;7</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M329.6995,-222.8796C329.8372,-214.6838 329.9834,-205.9891 330.126,-197.5013\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"333.6289,-197.3555 330.2975,-187.2981 326.6299,-197.2378 333.6289,-197.3555\"/>\n",
       "</g>\n",
       "<!-- 8 -->\n",
       "<g id=\"node9\" class=\"node\">\n",
       "<title>8</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M310,-68C310,-68 224,-68 224,-68 218,-68 212,-62 212,-56 212,-56 212,-12 212,-12 212,-6 218,0 224,0 224,0 310,0 310,0 316,0 322,-6 322,-12 322,-12 322,-56 322,-56 322,-62 316,-68 310,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"267\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"267\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 2</text>\n",
       "<text text-anchor=\"middle\" x=\"267\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 2, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"267\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 7&#45;&gt;8 -->\n",
       "<g id=\"edge8\" class=\"edge\">\n",
       "<title>7&#45;&gt;8</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M307.1688,-103.9815C302.1044,-95.1585 296.7475,-85.8258 291.6532,-76.9506\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"294.6807,-75.1942 286.667,-68.2637 288.6097,-78.6789 294.6807,-75.1942\"/>\n",
       "</g>\n",
       "<!-- 9 -->\n",
       "<g id=\"node10\" class=\"node\">\n",
       "<title>9</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M438,-68C438,-68 352,-68 352,-68 346,-68 340,-62 340,-56 340,-56 340,-12 340,-12 340,-6 346,0 352,0 352,0 438,0 438,0 444,0 450,-6 450,-12 450,-12 450,-56 450,-56 450,-62 444,-68 438,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"395\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"395\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 1</text>\n",
       "<text text-anchor=\"middle\" x=\"395\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [1, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"395\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 琴酒</text>\n",
       "</g>\n",
       "<!-- 7&#45;&gt;9 -->\n",
       "<g id=\"edge9\" class=\"edge\">\n",
       "<title>7&#45;&gt;9</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M354.8312,-103.9815C359.8956,-95.1585 365.2525,-85.8258 370.3468,-76.9506\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"373.3903,-78.6789 375.333,-68.2637 367.3193,-75.1942 373.3903,-78.6789\"/>\n",
       "</g>\n",
       "<!-- 11 -->\n",
       "<g id=\"node12\" class=\"node\">\n",
       "<title>11</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M511,-179.5C511,-179.5 425,-179.5 425,-179.5 419,-179.5 413,-173.5 413,-167.5 413,-167.5 413,-123.5 413,-123.5 413,-117.5 419,-111.5 425,-111.5 425,-111.5 511,-111.5 511,-111.5 517,-111.5 523,-117.5 523,-123.5 523,-123.5 523,-167.5 523,-167.5 523,-173.5 517,-179.5 511,-179.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"468\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"468\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 3</text>\n",
       "<text text-anchor=\"middle\" x=\"468\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 3, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"468\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 雪莉</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;11 -->\n",
       "<g id=\"edge11\" class=\"edge\">\n",
       "<title>10&#45;&gt;11</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M469.9507,-222.8796C469.6818,-212.2134 469.3916,-200.7021 469.1194,-189.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"472.6161,-189.7235 468.8651,-179.8149 465.6183,-189.9 472.6161,-189.7235\"/>\n",
       "</g>\n",
       "<!-- 12 -->\n",
       "<g id=\"node13\" class=\"node\">\n",
       "<title>12</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M646.5,-179.5C646.5,-179.5 553.5,-179.5 553.5,-179.5 547.5,-179.5 541.5,-173.5 541.5,-167.5 541.5,-167.5 541.5,-123.5 541.5,-123.5 541.5,-117.5 547.5,-111.5 553.5,-111.5 553.5,-111.5 646.5,-111.5 646.5,-111.5 652.5,-111.5 658.5,-117.5 658.5,-123.5 658.5,-123.5 658.5,-167.5 658.5,-167.5 658.5,-173.5 652.5,-179.5 646.5,-179.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"600\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"600\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 41</text>\n",
       "<text text-anchor=\"middle\" x=\"600\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [41, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"600\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 琴酒</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;12 -->\n",
       "<g id=\"edge12\" class=\"edge\">\n",
       "<title>10&#45;&gt;12</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M516.1179,-222.8796C528.7533,-211.2237 542.4826,-198.5587 555.105,-186.9148\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"557.8244,-189.168 562.8015,-179.8149 553.078,-184.0228 557.8244,-189.168\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x7f0fbb7af128>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "feature_names = ['酒精','苹果酸','灰','灰的碱性','镁','总酚',\n",
    "                 '类黄酮','非黄烷类酚类','花青素','颜色强度',\n",
    "                 '色调','od280/od315稀释葡萄酒','脯氨酸']\n",
    "dot_data = tree.export_graphviz(clf,\n",
    "                                out_file=None,\n",
    "                                feature_names=feature_names,\n",
    "                                class_names=['琴酒','雪莉','贝尔摩德'],\n",
    "                                filled=True,  # 设置叶颜色\n",
    "                                rounded=True,  # 设置叶子为圆形\n",
    "#                                 label=False\n",
    "                               )\n",
    "graph = graphviz.Source(dot_data)\n",
    "graph"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 探索决策树"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:18.001040Z",
     "start_time": "2019-12-06T05:07:17.988667Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('酒精', 0.02079843380404887),\n",
       " ('苹果酸', 0.0),\n",
       " ('灰', 0.0),\n",
       " ('灰的碱性', 0.0),\n",
       " ('镁', 0.0),\n",
       " ('总酚', 0.0),\n",
       " ('类黄酮', 0.4243707740617086),\n",
       " ('非黄烷类酚类', 0.0),\n",
       " ('花青素', 0.0),\n",
       " ('颜色强度', 0.2183301299894307),\n",
       " ('色调', 0.0),\n",
       " ('od280/od315稀释葡萄酒', 0.0),\n",
       " ('脯氨酸', 0.3365006621448118)]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 显示特征的重要性(权重)\n",
    "clf.feature_importances_\n",
    "[*zip(feature_names, clf.feature_importances_)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:19.281145Z",
     "start_time": "2019-12-06T05:07:19.275397Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9629629629629629"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = tree.DecisionTreeClassifier(criterion='entropy', random_state=30)\n",
    "clf = clf.fit(x_train, y_train)\n",
    "score = clf.score(x_test, y_test)\n",
    "score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 修改random_state&splitter参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:21.500936Z",
     "start_time": "2019-12-06T05:07:21.491283Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9444444444444444"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf = tree.DecisionTreeClassifier(criterion='entropy',\n",
    "                                  random_state=32,\n",
    "                                  splitter='random')  # splitter='best' by default\n",
    "clf = clf.fit(x_train, y_train)\n",
    "\n",
    "score = clf.score(x_test, y_test)\n",
    "score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:21.998047Z",
     "start_time": "2019-12-06T05:07:21.958419Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.40.1 (20161225.0304)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"850pt\" height=\"909pt\"\n",
       " viewBox=\"0.00 0.00 849.50 909.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 905)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"#ffffff\" stroke=\"transparent\" points=\"-4,4 -4,-905 845.5,-905 845.5,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\">\n",
       "<title>0</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.098039\" stroke=\"#000000\" d=\"M426,-901C426,-901 318,-901 318,-901 312,-901 306,-895 306,-889 306,-889 306,-830 306,-830 306,-824 312,-818 318,-818 318,-818 426,-818 426,-818 432,-818 438,-824 438,-830 438,-830 438,-889 438,-889 438,-895 432,-901 426,-901\"/>\n",
       "<text text-anchor=\"middle\" x=\"372\" y=\"-885.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">类黄酮 &lt;= 2.076</text>\n",
       "<text text-anchor=\"middle\" x=\"372\" y=\"-870.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 1.562</text>\n",
       "<text text-anchor=\"middle\" x=\"372\" y=\"-855.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 124</text>\n",
       "<text text-anchor=\"middle\" x=\"372\" y=\"-840.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 50, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"372\" y=\"-825.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\">\n",
       "<title>1</title>\n",
       "<path fill=\"#8139e5\" fill-opacity=\"0.188235\" stroke=\"#000000\" d=\"M349.5,-782C349.5,-782 248.5,-782 248.5,-782 242.5,-782 236.5,-776 236.5,-770 236.5,-770 236.5,-711 236.5,-711 236.5,-705 242.5,-699 248.5,-699 248.5,-699 349.5,-699 349.5,-699 355.5,-699 361.5,-705 361.5,-711 361.5,-711 361.5,-770 361.5,-770 361.5,-776 355.5,-782 349.5,-782\"/>\n",
       "<text text-anchor=\"middle\" x=\"299\" y=\"-766.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">色调 &lt;= 0.972</text>\n",
       "<text text-anchor=\"middle\" x=\"299\" y=\"-751.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.992</text>\n",
       "<text text-anchor=\"middle\" x=\"299\" y=\"-736.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 58</text>\n",
       "<text text-anchor=\"middle\" x=\"299\" y=\"-721.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 26, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"299\" y=\"-706.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\">\n",
       "<title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M346.4681,-817.8796C341.109,-809.1434 335.4021,-799.8404 329.8718,-790.8253\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"332.8533,-788.9919 324.6408,-782.2981 326.8865,-792.6522 332.8533,-788.9919\"/>\n",
       "<text text-anchor=\"middle\" x=\"318.8161\" y=\"-802.9101\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">True</text>\n",
       "</g>\n",
       "<!-- 8 -->\n",
       "<g id=\"node9\" class=\"node\">\n",
       "<title>8</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.427451\" stroke=\"#000000\" d=\"M500,-782C500,-782 392,-782 392,-782 386,-782 380,-776 380,-770 380,-770 380,-711 380,-711 380,-705 386,-699 392,-699 392,-699 500,-699 500,-699 506,-699 512,-705 512,-711 512,-711 512,-770 512,-770 512,-776 506,-782 500,-782\"/>\n",
       "<text text-anchor=\"middle\" x=\"446\" y=\"-766.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">脯氨酸 &lt;= 611.606</text>\n",
       "<text text-anchor=\"middle\" x=\"446\" y=\"-751.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.946</text>\n",
       "<text text-anchor=\"middle\" x=\"446\" y=\"-736.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 66</text>\n",
       "<text text-anchor=\"middle\" x=\"446\" y=\"-721.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 24, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"446\" y=\"-706.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;8 -->\n",
       "<g id=\"edge8\" class=\"edge\">\n",
       "<title>0&#45;&gt;8</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M397.8816,-817.8796C403.3142,-809.1434 409.0992,-799.8404 414.7053,-790.8253\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"417.6993,-792.6383 420.0079,-782.2981 411.7549,-788.9418 417.6993,-792.6383\"/>\n",
       "<text text-anchor=\"middle\" x=\"425.6847\" y=\"-802.945\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\">\n",
       "<title>2</title>\n",
       "<path fill=\"#8139e5\" fill-opacity=\"0.780392\" stroke=\"#000000\" d=\"M214.5,-663C214.5,-663 107.5,-663 107.5,-663 101.5,-663 95.5,-657 95.5,-651 95.5,-651 95.5,-592 95.5,-592 95.5,-586 101.5,-580 107.5,-580 107.5,-580 214.5,-580 214.5,-580 220.5,-580 226.5,-586 226.5,-592 226.5,-592 226.5,-651 226.5,-651 226.5,-657 220.5,-663 214.5,-663\"/>\n",
       "<text text-anchor=\"middle\" x=\"161\" y=\"-647.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 3.353</text>\n",
       "<text text-anchor=\"middle\" x=\"161\" y=\"-632.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.679</text>\n",
       "<text text-anchor=\"middle\" x=\"161\" y=\"-617.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 39</text>\n",
       "<text text-anchor=\"middle\" x=\"161\" y=\"-602.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 7, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"161\" y=\"-587.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\">\n",
       "<title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M250.7343,-698.8796C239.8722,-689.513 228.2544,-679.4948 217.1027,-669.8784\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"219.3306,-667.178 209.4717,-663.2981 214.7592,-672.4792 219.3306,-667.178\"/>\n",
       "</g>\n",
       "<!-- 7 -->\n",
       "<g id=\"node8\" class=\"node\">\n",
       "<title>7</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M349.5,-655.5C349.5,-655.5 256.5,-655.5 256.5,-655.5 250.5,-655.5 244.5,-649.5 244.5,-643.5 244.5,-643.5 244.5,-599.5 244.5,-599.5 244.5,-593.5 250.5,-587.5 256.5,-587.5 256.5,-587.5 349.5,-587.5 349.5,-587.5 355.5,-587.5 361.5,-593.5 361.5,-599.5 361.5,-599.5 361.5,-643.5 361.5,-643.5 361.5,-649.5 355.5,-655.5 349.5,-655.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"303\" y=\"-640.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"303\" y=\"-625.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 19</text>\n",
       "<text text-anchor=\"middle\" x=\"303\" y=\"-610.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 19, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"303\" y=\"-595.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;7 -->\n",
       "<g id=\"edge7\" class=\"edge\">\n",
       "<title>1&#45;&gt;7</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M300.399,-698.8796C300.7575,-688.2134 301.1445,-676.7021 301.5075,-665.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"305.0085,-665.9269 301.8466,-655.8149 298.0125,-665.6917 305.0085,-665.9269\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\">\n",
       "<title>3</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M98,-536.5C98,-536.5 12,-536.5 12,-536.5 6,-536.5 0,-530.5 0,-524.5 0,-524.5 0,-480.5 0,-480.5 0,-474.5 6,-468.5 12,-468.5 12,-468.5 98,-468.5 98,-468.5 104,-468.5 110,-474.5 110,-480.5 110,-480.5 110,-524.5 110,-524.5 110,-530.5 104,-536.5 98,-536.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-521.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-506.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 6</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-491.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 6, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"55\" y=\"-476.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\">\n",
       "<title>2&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M123.9263,-579.8796C113.7397,-568.4436 102.6878,-556.0363 92.4785,-544.575\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"94.8312,-541.9541 85.5662,-536.8149 89.6042,-546.6101 94.8312,-541.9541\"/>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\">\n",
       "<title>4</title>\n",
       "<path fill=\"#8139e5\" fill-opacity=\"0.968627\" stroke=\"#000000\" d=\"M248,-544C248,-544 140,-544 140,-544 134,-544 128,-538 128,-532 128,-532 128,-473 128,-473 128,-467 134,-461 140,-461 140,-461 248,-461 248,-461 254,-461 260,-467 260,-473 260,-473 260,-532 260,-532 260,-538 254,-544 248,-544\"/>\n",
       "<text text-anchor=\"middle\" x=\"194\" y=\"-528.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">脯氨酸 &lt;= 381.401</text>\n",
       "<text text-anchor=\"middle\" x=\"194\" y=\"-513.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.196</text>\n",
       "<text text-anchor=\"middle\" x=\"194\" y=\"-498.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 33</text>\n",
       "<text text-anchor=\"middle\" x=\"194\" y=\"-483.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 1, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"194\" y=\"-468.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\">\n",
       "<title>2&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M172.5418,-579.8796C174.8645,-571.5037 177.3318,-562.6067 179.7346,-553.942\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"183.1093,-554.8697 182.4089,-544.2981 176.3639,-552.9991 183.1093,-554.8697\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\">\n",
       "<title>5</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M139,-417.5C139,-417.5 53,-417.5 53,-417.5 47,-417.5 41,-411.5 41,-405.5 41,-405.5 41,-361.5 41,-361.5 41,-355.5 47,-349.5 53,-349.5 53,-349.5 139,-349.5 139,-349.5 145,-349.5 151,-355.5 151,-361.5 151,-361.5 151,-405.5 151,-405.5 151,-411.5 145,-417.5 139,-417.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"96\" y=\"-402.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"96\" y=\"-387.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 1</text>\n",
       "<text text-anchor=\"middle\" x=\"96\" y=\"-372.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"96\" y=\"-357.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\">\n",
       "<title>4&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M159.7244,-460.8796C150.3065,-449.4436 140.0887,-437.0363 130.65,-425.575\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"133.3182,-423.3093 124.2594,-417.8149 127.9147,-427.7592 133.3182,-423.3093\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\">\n",
       "<title>6</title>\n",
       "<path fill=\"#8139e5\" stroke=\"#000000\" d=\"M274.5,-417.5C274.5,-417.5 181.5,-417.5 181.5,-417.5 175.5,-417.5 169.5,-411.5 169.5,-405.5 169.5,-405.5 169.5,-361.5 169.5,-361.5 169.5,-355.5 175.5,-349.5 181.5,-349.5 181.5,-349.5 274.5,-349.5 274.5,-349.5 280.5,-349.5 286.5,-355.5 286.5,-361.5 286.5,-361.5 286.5,-405.5 286.5,-405.5 286.5,-411.5 280.5,-417.5 274.5,-417.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"228\" y=\"-402.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"228\" y=\"-387.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 32</text>\n",
       "<text text-anchor=\"middle\" x=\"228\" y=\"-372.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 0, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"228\" y=\"-357.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 4&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\">\n",
       "<title>4&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M205.8916,-460.8796C208.9705,-450.1034 212.2958,-438.4647 215.4092,-427.5677\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"218.8138,-428.3917 218.1957,-417.8149 212.0831,-426.4686 218.8138,-428.3917\"/>\n",
       "</g>\n",
       "<!-- 9 -->\n",
       "<g id=\"node10\" class=\"node\">\n",
       "<title>9</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M490.5,-655.5C490.5,-655.5 397.5,-655.5 397.5,-655.5 391.5,-655.5 385.5,-649.5 385.5,-643.5 385.5,-643.5 385.5,-599.5 385.5,-599.5 385.5,-593.5 391.5,-587.5 397.5,-587.5 397.5,-587.5 490.5,-587.5 490.5,-587.5 496.5,-587.5 502.5,-593.5 502.5,-599.5 502.5,-599.5 502.5,-643.5 502.5,-643.5 502.5,-649.5 496.5,-655.5 490.5,-655.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"444\" y=\"-640.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"444\" y=\"-625.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 17</text>\n",
       "<text text-anchor=\"middle\" x=\"444\" y=\"-610.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 17, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"444\" y=\"-595.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;9 -->\n",
       "<g id=\"edge9\" class=\"edge\">\n",
       "<title>8&#45;&gt;9</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M445.3005,-698.8796C445.1212,-688.2134 444.9278,-676.7021 444.7462,-665.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"448.2444,-665.7547 444.5767,-655.8149 441.2454,-665.8724 448.2444,-665.7547\"/>\n",
       "</g>\n",
       "<!-- 10 -->\n",
       "<g id=\"node11\" class=\"node\">\n",
       "<title>10</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.831373\" stroke=\"#000000\" d=\"M633,-663C633,-663 533,-663 533,-663 527,-663 521,-657 521,-651 521,-651 521,-592 521,-592 521,-586 527,-580 533,-580 533,-580 633,-580 633,-580 639,-580 645,-586 645,-592 645,-592 645,-651 645,-651 645,-657 639,-663 633,-663\"/>\n",
       "<text text-anchor=\"middle\" x=\"583\" y=\"-647.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 3.76</text>\n",
       "<text text-anchor=\"middle\" x=\"583\" y=\"-632.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.592</text>\n",
       "<text text-anchor=\"middle\" x=\"583\" y=\"-617.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 49</text>\n",
       "<text text-anchor=\"middle\" x=\"583\" y=\"-602.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 7, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"583\" y=\"-587.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;10 -->\n",
       "<g id=\"edge10\" class=\"edge\">\n",
       "<title>8&#45;&gt;10</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M493.916,-698.8796C504.6993,-689.513 516.2329,-679.4948 527.3039,-669.8784\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"529.6251,-672.4982 534.8795,-663.2981 525.0347,-667.2134 529.6251,-672.4982\"/>\n",
       "</g>\n",
       "<!-- 11 -->\n",
       "<g id=\"node12\" class=\"node\">\n",
       "<title>11</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.600000\" stroke=\"#000000\" d=\"M559.5,-544C559.5,-544 472.5,-544 472.5,-544 466.5,-544 460.5,-538 460.5,-532 460.5,-532 460.5,-473 460.5,-473 460.5,-467 466.5,-461 472.5,-461 472.5,-461 559.5,-461 559.5,-461 565.5,-461 571.5,-467 571.5,-473 571.5,-473 571.5,-532 571.5,-532 571.5,-538 565.5,-544 559.5,-544\"/>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-528.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">总酚 &lt;= 2.592</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-513.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.863</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-498.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 7</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-483.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [2, 5, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-468.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;11 -->\n",
       "<g id=\"edge11\" class=\"edge\">\n",
       "<title>10&#45;&gt;11</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M559.5667,-579.8796C554.6987,-571.2335 549.5181,-562.0322 544.4914,-553.1042\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"547.4893,-551.2947 539.5334,-544.2981 541.3897,-554.729 547.4893,-551.2947\"/>\n",
       "</g>\n",
       "<!-- 16 -->\n",
       "<g id=\"node17\" class=\"node\">\n",
       "<title>16</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.949020\" stroke=\"#000000\" d=\"M694.5,-544C694.5,-544 601.5,-544 601.5,-544 595.5,-544 589.5,-538 589.5,-532 589.5,-532 589.5,-473 589.5,-473 589.5,-467 595.5,-461 601.5,-461 601.5,-461 694.5,-461 694.5,-461 700.5,-461 706.5,-467 706.5,-473 706.5,-473 706.5,-532 706.5,-532 706.5,-538 700.5,-544 694.5,-544\"/>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-528.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">酒精 &lt;= 13.369</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-513.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.276</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-498.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 42</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-483.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [40, 2, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-468.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;16 -->\n",
       "<g id=\"edge16\" class=\"edge\">\n",
       "<title>10&#45;&gt;16</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M605.7338,-579.8796C610.4565,-571.2335 615.4824,-562.0322 620.359,-553.1042\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"623.447,-554.752 625.1691,-544.2981 617.3037,-551.3964 623.447,-554.752\"/>\n",
       "</g>\n",
       "<!-- 12 -->\n",
       "<g id=\"node13\" class=\"node\">\n",
       "<title>12</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.498039\" stroke=\"#000000\" d=\"M430.5,-425C430.5,-425 323.5,-425 323.5,-425 317.5,-425 311.5,-419 311.5,-413 311.5,-413 311.5,-354 311.5,-354 311.5,-348 317.5,-342 323.5,-342 323.5,-342 430.5,-342 430.5,-342 436.5,-342 442.5,-348 442.5,-354 442.5,-354 442.5,-413 442.5,-413 442.5,-419 436.5,-425 430.5,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"377\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 2.688</text>\n",
       "<text text-anchor=\"middle\" x=\"377\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.918</text>\n",
       "<text text-anchor=\"middle\" x=\"377\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 3</text>\n",
       "<text text-anchor=\"middle\" x=\"377\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [2, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"377\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 11&#45;&gt;12 -->\n",
       "<g id=\"edge12\" class=\"edge\">\n",
       "<title>11&#45;&gt;12</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M467.3845,-460.8796C456.4438,-451.513 444.7418,-441.4948 433.5092,-431.8784\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"435.6956,-429.1428 425.823,-425.2981 431.1432,-434.4603 435.6956,-429.1428\"/>\n",
       "</g>\n",
       "<!-- 15 -->\n",
       "<g id=\"node16\" class=\"node\">\n",
       "<title>15</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M559,-417.5C559,-417.5 473,-417.5 473,-417.5 467,-417.5 461,-411.5 461,-405.5 461,-405.5 461,-361.5 461,-361.5 461,-355.5 467,-349.5 473,-349.5 473,-349.5 559,-349.5 559,-349.5 565,-349.5 571,-355.5 571,-361.5 571,-361.5 571,-405.5 571,-405.5 571,-411.5 565,-417.5 559,-417.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-402.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-387.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 4</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-372.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 4, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"516\" y=\"-357.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 11&#45;&gt;15 -->\n",
       "<g id=\"edge15\" class=\"edge\">\n",
       "<title>11&#45;&gt;15</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M516,-460.8796C516,-450.2134 516,-438.7021 516,-427.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"519.5001,-427.8149 516,-417.8149 512.5001,-427.815 519.5001,-427.8149\"/>\n",
       "</g>\n",
       "<!-- 13 -->\n",
       "<g id=\"node14\" class=\"node\">\n",
       "<title>13</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M328,-298.5C328,-298.5 242,-298.5 242,-298.5 236,-298.5 230,-292.5 230,-286.5 230,-286.5 230,-242.5 230,-242.5 230,-236.5 236,-230.5 242,-230.5 242,-230.5 328,-230.5 328,-230.5 334,-230.5 340,-236.5 340,-242.5 340,-242.5 340,-286.5 340,-286.5 340,-292.5 334,-298.5 328,-298.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"285\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"285\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 1</text>\n",
       "<text text-anchor=\"middle\" x=\"285\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"285\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;13 -->\n",
       "<g id=\"edge13\" class=\"edge\">\n",
       "<title>12&#45;&gt;13</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M344.8229,-341.8796C336.0667,-330.5536 326.5738,-318.2748 317.7844,-306.9058\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"320.4146,-304.5856 311.5292,-298.8149 314.8766,-308.8671 320.4146,-304.5856\"/>\n",
       "</g>\n",
       "<!-- 14 -->\n",
       "<g id=\"node15\" class=\"node\">\n",
       "<title>14</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M456,-298.5C456,-298.5 370,-298.5 370,-298.5 364,-298.5 358,-292.5 358,-286.5 358,-286.5 358,-242.5 358,-242.5 358,-236.5 364,-230.5 370,-230.5 370,-230.5 456,-230.5 456,-230.5 462,-230.5 468,-236.5 468,-242.5 468,-242.5 468,-286.5 468,-286.5 468,-292.5 462,-298.5 456,-298.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"413\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"413\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 2</text>\n",
       "<text text-anchor=\"middle\" x=\"413\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [2, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"413\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 12&#45;&gt;14 -->\n",
       "<g id=\"edge14\" class=\"edge\">\n",
       "<title>12&#45;&gt;14</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M389.5911,-341.8796C392.8511,-331.1034 396.372,-319.4647 399.6686,-308.5677\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"403.0734,-309.4 402.619,-298.8149 396.3733,-307.373 403.0734,-309.4\"/>\n",
       "</g>\n",
       "<!-- 17 -->\n",
       "<g id=\"node18\" class=\"node\">\n",
       "<title>17</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.819608\" stroke=\"#000000\" d=\"M694.5,-425C694.5,-425 601.5,-425 601.5,-425 595.5,-425 589.5,-419 589.5,-413 589.5,-413 589.5,-354 589.5,-354 589.5,-348 595.5,-342 601.5,-342 601.5,-342 694.5,-342 694.5,-342 700.5,-342 706.5,-348 706.5,-354 706.5,-354 706.5,-413 706.5,-413 706.5,-419 700.5,-425 694.5,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">酒精 &lt;= 12.654</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.619</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 13</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [11, 2, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"648\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 16&#45;&gt;17 -->\n",
       "<g id=\"edge17\" class=\"edge\">\n",
       "<title>16&#45;&gt;17</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M648,-460.8796C648,-452.6838 648,-443.9891 648,-435.5013\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"651.5001,-435.298 648,-425.2981 644.5001,-435.2981 651.5001,-435.298\"/>\n",
       "</g>\n",
       "<!-- 24 -->\n",
       "<g id=\"node25\" class=\"node\">\n",
       "<title>24</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M829.5,-417.5C829.5,-417.5 736.5,-417.5 736.5,-417.5 730.5,-417.5 724.5,-411.5 724.5,-405.5 724.5,-405.5 724.5,-361.5 724.5,-361.5 724.5,-355.5 730.5,-349.5 736.5,-349.5 736.5,-349.5 829.5,-349.5 829.5,-349.5 835.5,-349.5 841.5,-355.5 841.5,-361.5 841.5,-361.5 841.5,-405.5 841.5,-405.5 841.5,-411.5 835.5,-417.5 829.5,-417.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"783\" y=\"-402.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"783\" y=\"-387.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 29</text>\n",
       "<text text-anchor=\"middle\" x=\"783\" y=\"-372.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [29, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"783\" y=\"-357.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 16&#45;&gt;24 -->\n",
       "<g id=\"edge24\" class=\"edge\">\n",
       "<title>16&#45;&gt;24</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M695.2164,-460.8796C708.5642,-449.1138 723.0785,-436.3197 736.3904,-424.5855\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"738.884,-427.0531 744.0713,-417.8149 734.2552,-421.8019 738.884,-427.0531\"/>\n",
       "</g>\n",
       "<!-- 18 -->\n",
       "<g id=\"node19\" class=\"node\">\n",
       "<title>18</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M621,-298.5C621,-298.5 535,-298.5 535,-298.5 529,-298.5 523,-292.5 523,-286.5 523,-286.5 523,-242.5 523,-242.5 523,-236.5 529,-230.5 535,-230.5 535,-230.5 621,-230.5 621,-230.5 627,-230.5 633,-236.5 633,-242.5 633,-242.5 633,-286.5 633,-286.5 633,-292.5 627,-298.5 621,-298.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"578\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"578\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 1</text>\n",
       "<text text-anchor=\"middle\" x=\"578\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"578\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 17&#45;&gt;18 -->\n",
       "<g id=\"edge18\" class=\"edge\">\n",
       "<title>17&#45;&gt;18</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M623.5174,-341.8796C616.9844,-330.7735 609.9125,-318.7513 603.3348,-307.5691\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"606.2723,-305.6597 598.1853,-298.8149 600.2387,-309.2089 606.2723,-305.6597\"/>\n",
       "</g>\n",
       "<!-- 19 -->\n",
       "<g id=\"node20\" class=\"node\">\n",
       "<title>19</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.909804\" stroke=\"#000000\" d=\"M770.5,-306C770.5,-306 663.5,-306 663.5,-306 657.5,-306 651.5,-300 651.5,-294 651.5,-294 651.5,-235 651.5,-235 651.5,-229 657.5,-223 663.5,-223 663.5,-223 770.5,-223 770.5,-223 776.5,-223 782.5,-229 782.5,-235 782.5,-235 782.5,-294 782.5,-294 782.5,-300 776.5,-306 770.5,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"717\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 4.209</text>\n",
       "<text text-anchor=\"middle\" x=\"717\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.414</text>\n",
       "<text text-anchor=\"middle\" x=\"717\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 12</text>\n",
       "<text text-anchor=\"middle\" x=\"717\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [11, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"717\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 17&#45;&gt;19 -->\n",
       "<g id=\"edge19\" class=\"edge\">\n",
       "<title>17&#45;&gt;19</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M672.1329,-341.8796C677.1461,-333.2335 682.4813,-324.0322 687.6581,-315.1042\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"690.7758,-316.7047 692.7641,-306.2981 684.7202,-313.1934 690.7758,-316.7047\"/>\n",
       "</g>\n",
       "<!-- 20 -->\n",
       "<g id=\"node21\" class=\"node\">\n",
       "<title>20</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.498039\" stroke=\"#000000\" d=\"M695.5,-187C695.5,-187 608.5,-187 608.5,-187 602.5,-187 596.5,-181 596.5,-175 596.5,-175 596.5,-116 596.5,-116 596.5,-110 602.5,-104 608.5,-104 608.5,-104 695.5,-104 695.5,-104 701.5,-104 707.5,-110 707.5,-116 707.5,-116 707.5,-175 707.5,-175 707.5,-181 701.5,-187 695.5,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"652\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">镁 &lt;= 87.444</text>\n",
       "<text text-anchor=\"middle\" x=\"652\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.918</text>\n",
       "<text text-anchor=\"middle\" x=\"652\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 3</text>\n",
       "<text text-anchor=\"middle\" x=\"652\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [2, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"652\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 19&#45;&gt;20 -->\n",
       "<g id=\"edge20\" class=\"edge\">\n",
       "<title>19&#45;&gt;20</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M694.2662,-222.8796C689.5435,-214.2335 684.5176,-205.0322 679.641,-196.1042\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"682.6963,-194.3964 674.8309,-187.2981 676.553,-197.752 682.6963,-194.3964\"/>\n",
       "</g>\n",
       "<!-- 23 -->\n",
       "<g id=\"node24\" class=\"node\">\n",
       "<title>23</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M824,-179.5C824,-179.5 738,-179.5 738,-179.5 732,-179.5 726,-173.5 726,-167.5 726,-167.5 726,-123.5 726,-123.5 726,-117.5 732,-111.5 738,-111.5 738,-111.5 824,-111.5 824,-111.5 830,-111.5 836,-117.5 836,-123.5 836,-123.5 836,-167.5 836,-167.5 836,-173.5 830,-179.5 824,-179.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"781\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"781\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 9</text>\n",
       "<text text-anchor=\"middle\" x=\"781\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [9, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"781\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 19&#45;&gt;23 -->\n",
       "<g id=\"edge23\" class=\"edge\">\n",
       "<title>19&#45;&gt;23</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M739.3841,-222.8796C745.298,-211.8835 751.6949,-199.9893 757.658,-188.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"760.8908,-190.2799 762.5449,-179.8149 754.7258,-186.9642 760.8908,-190.2799\"/>\n",
       "</g>\n",
       "<!-- 21 -->\n",
       "<g id=\"node22\" class=\"node\">\n",
       "<title>21</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M631,-68C631,-68 545,-68 545,-68 539,-68 533,-62 533,-56 533,-56 533,-12 533,-12 533,-6 539,0 545,0 545,0 631,0 631,0 637,0 643,-6 643,-12 643,-12 643,-56 643,-56 643,-62 637,-68 631,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"588\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"588\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 1</text>\n",
       "<text text-anchor=\"middle\" x=\"588\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 1, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"588\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 20&#45;&gt;21 -->\n",
       "<g id=\"edge21\" class=\"edge\">\n",
       "<title>20&#45;&gt;21</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M628.1688,-103.9815C623.1044,-95.1585 617.7475,-85.8258 612.6532,-76.9506\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"615.6807,-75.1942 607.667,-68.2637 609.6097,-78.6789 615.6807,-75.1942\"/>\n",
       "</g>\n",
       "<!-- 22 -->\n",
       "<g id=\"node23\" class=\"node\">\n",
       "<title>22</title>\n",
       "<path fill=\"#e58139\" stroke=\"#000000\" d=\"M759,-68C759,-68 673,-68 673,-68 667,-68 661,-62 661,-56 661,-56 661,-12 661,-12 661,-6 667,0 673,0 673,0 759,0 759,0 765,0 771,-6 771,-12 771,-12 771,-56 771,-56 771,-62 765,-68 759,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"716\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">entropy = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"716\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 2</text>\n",
       "<text text-anchor=\"middle\" x=\"716\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [2, 0, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"716\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 清酒</text>\n",
       "</g>\n",
       "<!-- 20&#45;&gt;22 -->\n",
       "<g id=\"edge22\" class=\"edge\">\n",
       "<title>20&#45;&gt;22</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M675.8312,-103.9815C680.8956,-95.1585 686.2525,-85.8258 691.3468,-76.9506\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"694.3903,-78.6789 696.333,-68.2637 688.3193,-75.1942 694.3903,-78.6789\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x7f0fbb74e550>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dot_data = tree.export_graphviz(clf, \n",
    "                                feature_names=feature_names,\n",
    "                                class_names=['清酒', '茅台','五粮液'],\n",
    "                                filled=True,\n",
    "                                rounded=True\n",
    "                                )\n",
    "graphviz.Source(dot_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:22.695741Z",
     "start_time": "2019-12-06T05:07:22.684002Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('酒精', 0.034259741095031854),\n",
       " ('苹果酸', 0.0),\n",
       " ('灰', 0.0),\n",
       " ('灰的碱性', 0.0),\n",
       " ('镁', 0.01422610193269686),\n",
       " ('总酚', 0.01697367963840301),\n",
       " ('类黄酮', 0.3805064782807622),\n",
       " ('非黄烷类酚类', 0.0),\n",
       " ('花青素', 0.0),\n",
       " ('颜色强度', 0.1876058600625648),\n",
       " ('色调', 0.16045561928227117),\n",
       " ('od280/od315稀释葡萄酒', 0.0),\n",
       " ('脯氨酸', 0.20597251970827027)]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "[*zip(feature_names, clf.feature_importances_)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 剪枝参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:23.970797Z",
     "start_time": "2019-12-06T05:07:23.959424Z"
    }
   },
   "outputs": [],
   "source": [
    "def draw_graph(**kwargs):\n",
    "    clf=tree.DecisionTreeClassifier(**kwargs)\n",
    "    clf = clf.fit(x_train, y_train)\n",
    "    dot_data = tree.export_graphviz(clf, \n",
    "                                    feature_names=feature_names,\n",
    "                                    class_names=['茅台', '五粮液', '泸州老窖'],\n",
    "                                    filled=True,\n",
    "                                    rounded=True)\n",
    "    graph = graphviz.Source(dot_data)\n",
    "    return graph, clf.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:25.000125Z",
     "start_time": "2019-12-06T05:07:24.957274Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\n",
       "<!-- Generated by graphviz version 2.40.1 (20161225.0304)\n",
       " -->\n",
       "<!-- Title: Tree Pages: 1 -->\n",
       "<svg width=\"813pt\" height=\"433pt\"\n",
       " viewBox=\"0.00 0.00 813.00 433.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 429)\">\n",
       "<title>Tree</title>\n",
       "<polygon fill=\"#ffffff\" stroke=\"transparent\" points=\"-4,4 -4,-429 809,-429 809,4 -4,4\"/>\n",
       "<!-- 0 -->\n",
       "<g id=\"node1\" class=\"node\">\n",
       "<title>0</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.098039\" stroke=\"#000000\" d=\"M460.5,-425C460.5,-425 352.5,-425 352.5,-425 346.5,-425 340.5,-419 340.5,-413 340.5,-413 340.5,-354 340.5,-354 340.5,-348 346.5,-342 352.5,-342 352.5,-342 460.5,-342 460.5,-342 466.5,-342 472.5,-348 472.5,-354 472.5,-354 472.5,-413 472.5,-413 472.5,-419 466.5,-425 460.5,-425\"/>\n",
       "<text text-anchor=\"middle\" x=\"406.5\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 4.26</text>\n",
       "<text text-anchor=\"middle\" x=\"406.5\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.656</text>\n",
       "<text text-anchor=\"middle\" x=\"406.5\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 124</text>\n",
       "<text text-anchor=\"middle\" x=\"406.5\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [42, 50, 32]</text>\n",
       "<text text-anchor=\"middle\" x=\"406.5\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 1 -->\n",
       "<g id=\"node2\" class=\"node\">\n",
       "<title>1</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.831373\" stroke=\"#000000\" d=\"M362,-306C362,-306 255,-306 255,-306 249,-306 243,-300 243,-294 243,-294 243,-235 243,-235 243,-229 249,-223 255,-223 255,-223 362,-223 362,-223 368,-223 374,-229 374,-235 374,-235 374,-294 374,-294 374,-300 368,-306 362,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"308.5\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">颜色强度 &lt;= 3.543</text>\n",
       "<text text-anchor=\"middle\" x=\"308.5\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.265</text>\n",
       "<text text-anchor=\"middle\" x=\"308.5\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 53</text>\n",
       "<text text-anchor=\"middle\" x=\"308.5\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [6, 45, 2]</text>\n",
       "<text text-anchor=\"middle\" x=\"308.5\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;1 -->\n",
       "<g id=\"edge1\" class=\"edge\">\n",
       "<title>0&#45;&gt;1</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M372.2244,-341.8796C364.8816,-332.9633 357.0524,-323.4565 349.4854,-314.268\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"351.9808,-311.7924 342.9219,-306.2981 346.5773,-316.2424 351.9808,-311.7924\"/>\n",
       "<text text-anchor=\"middle\" x=\"340.5138\" y=\"-327.4818\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">True</text>\n",
       "</g>\n",
       "<!-- 8 -->\n",
       "<g id=\"node9\" class=\"node\">\n",
       "<title>8</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.145098\" stroke=\"#000000\" d=\"M604.5,-306C604.5,-306 404.5,-306 404.5,-306 398.5,-306 392.5,-300 392.5,-294 392.5,-294 392.5,-235 392.5,-235 392.5,-229 398.5,-223 404.5,-223 404.5,-223 604.5,-223 604.5,-223 610.5,-223 616.5,-229 616.5,-235 616.5,-235 616.5,-294 616.5,-294 616.5,-300 610.5,-306 604.5,-306\"/>\n",
       "<text text-anchor=\"middle\" x=\"504.5\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">od280/od315稀释葡萄酒 &lt;= 1.809</text>\n",
       "<text text-anchor=\"middle\" x=\"504.5\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.559</text>\n",
       "<text text-anchor=\"middle\" x=\"504.5\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 71</text>\n",
       "<text text-anchor=\"middle\" x=\"504.5\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [36, 5, 30]</text>\n",
       "<text text-anchor=\"middle\" x=\"504.5\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 0&#45;&gt;8 -->\n",
       "<g id=\"edge8\" class=\"edge\">\n",
       "<title>0&#45;&gt;8</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M440.7756,-341.8796C448.1184,-332.9633 455.9476,-323.4565 463.5146,-314.268\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"466.4227,-316.2424 470.0781,-306.2981 461.0192,-311.7924 466.4227,-316.2424\"/>\n",
       "<text text-anchor=\"middle\" x=\"472.4862\" y=\"-327.4818\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">False</text>\n",
       "</g>\n",
       "<!-- 2 -->\n",
       "<g id=\"node3\" class=\"node\">\n",
       "<title>2</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.972549\" stroke=\"#000000\" d=\"M235,-187C235,-187 142,-187 142,-187 136,-187 130,-181 130,-175 130,-175 130,-116 130,-116 130,-110 136,-104 142,-104 142,-104 235,-104 235,-104 241,-104 247,-110 247,-116 247,-116 247,-175 247,-175 247,-181 241,-187 235,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"188.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">酒精 &lt;= 12.904</text>\n",
       "<text text-anchor=\"middle\" x=\"188.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.049</text>\n",
       "<text text-anchor=\"middle\" x=\"188.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 40</text>\n",
       "<text text-anchor=\"middle\" x=\"188.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [1, 39, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"188.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;2 -->\n",
       "<g id=\"edge2\" class=\"edge\">\n",
       "<title>1&#45;&gt;2</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M266.5298,-222.8796C257.2662,-213.6931 247.3704,-203.8798 237.8448,-194.4336\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"240.2144,-191.8543 230.6493,-187.2981 235.2854,-196.8247 240.2144,-191.8543\"/>\n",
       "</g>\n",
       "<!-- 5 -->\n",
       "<g id=\"node6\" class=\"node\">\n",
       "<title>5</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.125490\" stroke=\"#000000\" d=\"M370,-187C370,-187 277,-187 277,-187 271,-187 265,-181 265,-175 265,-175 265,-116 265,-116 265,-110 271,-104 277,-104 277,-104 370,-104 370,-104 376,-104 382,-110 382,-116 382,-116 382,-175 382,-175 382,-181 376,-187 370,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"323.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">苹果酸 &lt;= 3.523</text>\n",
       "<text text-anchor=\"middle\" x=\"323.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.615</text>\n",
       "<text text-anchor=\"middle\" x=\"323.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 13</text>\n",
       "<text text-anchor=\"middle\" x=\"323.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [5, 6, 2]</text>\n",
       "<text text-anchor=\"middle\" x=\"323.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 1&#45;&gt;5 -->\n",
       "<g id=\"edge5\" class=\"edge\">\n",
       "<title>1&#45;&gt;5</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M313.7463,-222.8796C314.7907,-214.5938 315.8994,-205.798 316.9805,-197.2216\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"320.4532,-197.6573 318.2313,-187.2981 313.5081,-196.7818 320.4532,-197.6573\"/>\n",
       "</g>\n",
       "<!-- 3 -->\n",
       "<g id=\"node4\" class=\"node\">\n",
       "<title>3</title>\n",
       "<path fill=\"#39e581\" stroke=\"#000000\" d=\"M105,-68C105,-68 12,-68 12,-68 6,-68 0,-62 0,-56 0,-56 0,-12 0,-12 0,-6 6,0 12,0 12,0 105,0 105,0 111,0 117,-6 117,-12 117,-12 117,-56 117,-56 117,-62 111,-68 105,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"58.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"58.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 37</text>\n",
       "<text text-anchor=\"middle\" x=\"58.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 37, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"58.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;3 -->\n",
       "<g id=\"edge3\" class=\"edge\">\n",
       "<title>2&#45;&gt;3</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M140.0928,-103.9815C129.0558,-94.5151 117.3347,-84.462 106.3223,-75.0168\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"108.3178,-72.1173 98.4487,-68.2637 103.7606,-77.4307 108.3178,-72.1173\"/>\n",
       "</g>\n",
       "<!-- 4 -->\n",
       "<g id=\"node5\" class=\"node\">\n",
       "<title>4</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.498039\" stroke=\"#000000\" d=\"M233.5,-68C233.5,-68 147.5,-68 147.5,-68 141.5,-68 135.5,-62 135.5,-56 135.5,-56 135.5,-12 135.5,-12 135.5,-6 141.5,0 147.5,0 147.5,0 233.5,0 233.5,0 239.5,0 245.5,-6 245.5,-12 245.5,-12 245.5,-56 245.5,-56 245.5,-62 239.5,-68 233.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"190.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.444</text>\n",
       "<text text-anchor=\"middle\" x=\"190.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 3</text>\n",
       "<text text-anchor=\"middle\" x=\"190.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [1, 2, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"190.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 2&#45;&gt;4 -->\n",
       "<g id=\"edge4\" class=\"edge\">\n",
       "<title>2&#45;&gt;4</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M189.2447,-103.9815C189.3947,-95.618 189.553,-86.7965 189.7046,-78.3409\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"193.2054,-78.3249 189.8854,-68.2637 186.2065,-78.1993 193.2054,-78.3249\"/>\n",
       "</g>\n",
       "<!-- 6 -->\n",
       "<g id=\"node7\" class=\"node\">\n",
       "<title>6</title>\n",
       "<path fill=\"#39e581\" fill-opacity=\"0.164706\" stroke=\"#000000\" d=\"M364.5,-68C364.5,-68 278.5,-68 278.5,-68 272.5,-68 266.5,-62 266.5,-56 266.5,-56 266.5,-12 266.5,-12 266.5,-6 272.5,0 278.5,0 278.5,0 364.5,0 364.5,0 370.5,0 376.5,-6 376.5,-12 376.5,-12 376.5,-56 376.5,-56 376.5,-62 370.5,-68 364.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"321.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.496</text>\n",
       "<text text-anchor=\"middle\" x=\"321.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 11</text>\n",
       "<text text-anchor=\"middle\" x=\"321.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [5, 6, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"321.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 五粮液</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;6 -->\n",
       "<g id=\"edge6\" class=\"edge\">\n",
       "<title>5&#45;&gt;6</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M322.7553,-103.9815C322.6053,-95.618 322.447,-86.7965 322.2954,-78.3409\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"325.7935,-78.1993 322.1146,-68.2637 318.7946,-78.3249 325.7935,-78.1993\"/>\n",
       "</g>\n",
       "<!-- 7 -->\n",
       "<g id=\"node8\" class=\"node\">\n",
       "<title>7</title>\n",
       "<path fill=\"#8139e5\" stroke=\"#000000\" d=\"M502.5,-68C502.5,-68 406.5,-68 406.5,-68 400.5,-68 394.5,-62 394.5,-56 394.5,-56 394.5,-12 394.5,-12 394.5,-6 400.5,0 406.5,0 406.5,0 502.5,0 502.5,0 508.5,0 514.5,-6 514.5,-12 514.5,-12 514.5,-56 514.5,-56 514.5,-62 508.5,-68 502.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"454.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"454.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 2</text>\n",
       "<text text-anchor=\"middle\" x=\"454.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 0, 2]</text>\n",
       "<text text-anchor=\"middle\" x=\"454.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 泸州老窖</text>\n",
       "</g>\n",
       "<!-- 5&#45;&gt;7 -->\n",
       "<g id=\"edge7\" class=\"edge\">\n",
       "<title>5&#45;&gt;7</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M372.2796,-103.9815C383.4015,-94.5151 395.2128,-84.462 406.3099,-75.0168\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"408.8975,-77.4105 414.244,-68.2637 404.3604,-72.08 408.8975,-77.4105\"/>\n",
       "</g>\n",
       "<!-- 9 -->\n",
       "<g id=\"node10\" class=\"node\">\n",
       "<title>9</title>\n",
       "<path fill=\"#8139e5\" stroke=\"#000000\" d=\"M537.5,-179.5C537.5,-179.5 441.5,-179.5 441.5,-179.5 435.5,-179.5 429.5,-173.5 429.5,-167.5 429.5,-167.5 429.5,-123.5 429.5,-123.5 429.5,-117.5 435.5,-111.5 441.5,-111.5 441.5,-111.5 537.5,-111.5 537.5,-111.5 543.5,-111.5 549.5,-117.5 549.5,-123.5 549.5,-123.5 549.5,-167.5 549.5,-167.5 549.5,-173.5 543.5,-179.5 537.5,-179.5\"/>\n",
       "<text text-anchor=\"middle\" x=\"489.5\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"489.5\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 22</text>\n",
       "<text text-anchor=\"middle\" x=\"489.5\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 0, 22]</text>\n",
       "<text text-anchor=\"middle\" x=\"489.5\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 泸州老窖</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;9 -->\n",
       "<g id=\"edge9\" class=\"edge\">\n",
       "<title>8&#45;&gt;9</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M499.2537,-222.8796C497.9092,-212.2134 496.4582,-200.7021 495.0968,-189.9015\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"498.5486,-189.2987 493.8254,-179.8149 491.6036,-190.1742 498.5486,-189.2987\"/>\n",
       "</g>\n",
       "<!-- 10 -->\n",
       "<g id=\"node11\" class=\"node\">\n",
       "<title>10</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.682353\" stroke=\"#000000\" d=\"M673,-187C673,-187 580,-187 580,-187 574,-187 568,-181 568,-175 568,-175 568,-116 568,-116 568,-110 574,-104 580,-104 580,-104 673,-104 673,-104 679,-104 685,-110 685,-116 685,-116 685,-175 685,-175 685,-181 679,-187 673,-187\"/>\n",
       "<text text-anchor=\"middle\" x=\"626.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">类黄酮 &lt;= 1.397</text>\n",
       "<text text-anchor=\"middle\" x=\"626.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.423</text>\n",
       "<text text-anchor=\"middle\" x=\"626.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 49</text>\n",
       "<text text-anchor=\"middle\" x=\"626.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [36, 5, 8]</text>\n",
       "<text text-anchor=\"middle\" x=\"626.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 8&#45;&gt;10 -->\n",
       "<g id=\"edge10\" class=\"edge\">\n",
       "<title>8&#45;&gt;10</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M547.1697,-222.8796C556.5877,-213.6931 566.6484,-203.8798 576.3328,-194.4336\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"578.9335,-196.7861 583.6482,-187.2981 574.0457,-191.7751 578.9335,-196.7861\"/>\n",
       "</g>\n",
       "<!-- 11 -->\n",
       "<g id=\"node12\" class=\"node\">\n",
       "<title>11</title>\n",
       "<path fill=\"#8139e5\" stroke=\"#000000\" d=\"M657.5,-68C657.5,-68 561.5,-68 561.5,-68 555.5,-68 549.5,-62 549.5,-56 549.5,-56 549.5,-12 549.5,-12 549.5,-6 555.5,0 561.5,0 561.5,0 657.5,0 657.5,0 663.5,0 669.5,-6 669.5,-12 669.5,-12 669.5,-56 669.5,-56 669.5,-62 663.5,-68 657.5,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"609.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.0</text>\n",
       "<text text-anchor=\"middle\" x=\"609.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 8</text>\n",
       "<text text-anchor=\"middle\" x=\"609.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [0, 0, 8]</text>\n",
       "<text text-anchor=\"middle\" x=\"609.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 泸州老窖</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;11 -->\n",
       "<g id=\"edge11\" class=\"edge\">\n",
       "<title>10&#45;&gt;11</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M620.1698,-103.9815C618.8947,-95.618 617.5497,-86.7965 616.2605,-78.3409\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"619.6914,-77.6219 614.7241,-68.2637 612.7714,-78.677 619.6914,-77.6219\"/>\n",
       "</g>\n",
       "<!-- 12 -->\n",
       "<g id=\"node13\" class=\"node\">\n",
       "<title>12</title>\n",
       "<path fill=\"#e58139\" fill-opacity=\"0.862745\" stroke=\"#000000\" d=\"M793,-68C793,-68 700,-68 700,-68 694,-68 688,-62 688,-56 688,-56 688,-12 688,-12 688,-6 694,0 700,0 700,0 793,0 793,0 799,0 805,-6 805,-12 805,-12 805,-56 805,-56 805,-62 799,-68 793,-68\"/>\n",
       "<text text-anchor=\"middle\" x=\"746.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">gini = 0.214</text>\n",
       "<text text-anchor=\"middle\" x=\"746.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">samples = 41</text>\n",
       "<text text-anchor=\"middle\" x=\"746.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">value = [36, 5, 0]</text>\n",
       "<text text-anchor=\"middle\" x=\"746.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\" fill=\"#000000\">class = 茅台</text>\n",
       "</g>\n",
       "<!-- 10&#45;&gt;12 -->\n",
       "<g id=\"edge12\" class=\"edge\">\n",
       "<title>10&#45;&gt;12</title>\n",
       "<path fill=\"none\" stroke=\"#000000\" d=\"M671.1836,-103.9815C681.2727,-94.607 691.981,-84.6572 702.0601,-75.2921\"/>\n",
       "<polygon fill=\"#000000\" stroke=\"#000000\" points=\"704.6809,-77.6346 709.6243,-68.2637 699.9161,-72.5065 704.6809,-77.6346\"/>\n",
       "</g>\n",
       "</g>\n",
       "</svg>\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x7f0fbb74eb00>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph, score = draw_graph(random_state=30,\n",
    "                          splitter='random',\n",
    "                          max_depth=3,\n",
    "                          \n",
    "                          )\n",
    "graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-06T05:07:25.524703Z",
     "start_time": "2019-12-06T05:07:25.521289Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8518518518518519"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:52:16.321622Z",
     "start_time": "2019-12-05T00:52:16.037859Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\r\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\r\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\r\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\r\n",
       " -->\r\n",
       "<!-- Title: Tree Pages: 1 -->\r\n",
       "<svg width=\"704pt\" height=\"433pt\"\r\n",
       " viewBox=\"0.00 0.00 704.00 433.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\r\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 429)\">\r\n",
       "<title>Tree</title>\r\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-429 700,-429 700,4 -4,4\"/>\r\n",
       "<!-- 0 -->\r\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\r\n",
       "<path fill=\"#d8fae6\" stroke=\"black\" d=\"M333,-425C333,-425 220,-425 220,-425 214,-425 208,-419 208,-413 208,-413 208,-354 208,-354 208,-348 214,-342 220,-342 220,-342 333,-342 333,-342 339,-342 345,-348 345,-354 345,-354 345,-413 345,-413 345,-419 339,-425 333,-425\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">颜色强度 &lt;= 3.956</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.65</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 124</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [37, 54, 33]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1 -->\r\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\r\n",
       "<path fill=\"#52e891\" stroke=\"black\" d=\"M232.5,-306C232.5,-306 128.5,-306 128.5,-306 122.5,-306 116.5,-300 116.5,-294 116.5,-294 116.5,-235 116.5,-235 116.5,-229 122.5,-223 128.5,-223 128.5,-223 232.5,-223 232.5,-223 238.5,-223 244.5,-229 244.5,-235 244.5,-235 244.5,-294 244.5,-294 244.5,-300 238.5,-306 232.5,-306\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">颜色强度 &lt;= 3.314</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.208</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 52</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [5, 46, 1]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;1 -->\r\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M243.193,-341.907C235.747,-332.832 227.779,-323.121 220.106,-313.769\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"222.797,-311.531 213.748,-306.021 217.385,-315.972 222.797,-311.531\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"211.295\" y=\"-327.2\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">True</text>\r\n",
       "</g>\r\n",
       "<!-- 4 -->\r\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\r\n",
       "<path fill=\"#ffffff\" stroke=\"black\" d=\"M470,-306C470,-306 275,-306 275,-306 269,-306 263,-300 263,-294 263,-294 263,-235 263,-235 263,-229 269,-223 275,-223 275,-223 470,-223 470,-223 476,-223 482,-229 482,-235 482,-235 482,-294 482,-294 482,-300 476,-306 470,-306\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">od280/od315稀释葡萄酒 &lt;= 2.11</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.593</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 72</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [32, 8, 32]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;4 -->\r\n",
       "<g id=\"edge4\" class=\"edge\"><title>0&#45;&gt;4</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M309.807,-341.907C317.253,-332.832 325.221,-323.121 332.894,-313.769\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"335.615,-315.972 339.252,-306.021 330.203,-311.531 335.615,-315.972\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"341.705\" y=\"-327.2\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">False</text>\r\n",
       "</g>\r\n",
       "<!-- 2 -->\r\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\r\n",
       "<path fill=\"#39e581\" stroke=\"black\" d=\"M109,-179.5C109,-179.5 12,-179.5 12,-179.5 6,-179.5 0,-173.5 0,-167.5 0,-167.5 0,-123.5 0,-123.5 0,-117.5 6,-111.5 12,-111.5 12,-111.5 109,-111.5 109,-111.5 115,-111.5 121,-117.5 121,-123.5 121,-123.5 121,-167.5 121,-167.5 121,-173.5 115,-179.5 109,-179.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 38</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 38, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;2 -->\r\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M138.867,-222.907C126.87,-211.211 113.79,-198.457 101.813,-186.78\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"104.121,-184.142 94.5174,-179.667 99.234,-189.154 104.121,-184.142\"/>\r\n",
       "</g>\r\n",
       "<!-- 3 -->\r\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\r\n",
       "<path fill=\"#bdf6d5\" stroke=\"black\" d=\"M240,-179.5C240,-179.5 151,-179.5 151,-179.5 145,-179.5 139,-173.5 139,-167.5 139,-167.5 139,-123.5 139,-123.5 139,-117.5 145,-111.5 151,-111.5 151,-111.5 240,-111.5 240,-111.5 246,-111.5 252,-117.5 252,-123.5 252,-123.5 252,-167.5 252,-167.5 252,-173.5 246,-179.5 240,-179.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.541</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 14</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [5, 8, 1]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;3 -->\r\n",
       "<g id=\"edge3\" class=\"edge\"><title>1&#45;&gt;3</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M185.704,-222.907C187.076,-212.204 188.562,-200.615 189.952,-189.776\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"193.448,-190.031 191.248,-179.667 186.505,-189.141 193.448,-190.031\"/>\r\n",
       "</g>\r\n",
       "<!-- 5 -->\r\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\r\n",
       "<path fill=\"#8946e7\" stroke=\"black\" d=\"M407,-187C407,-187 310,-187 310,-187 304,-187 298,-181 298,-175 298,-175 298,-116 298,-116 298,-110 304,-104 310,-104 310,-104 407,-104 407,-104 413,-104 419,-110 419,-116 419,-116 419,-175 419,-175 419,-181 413,-187 407,-187\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">灰 &lt;= 2.448</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.117</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 32</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 2, 30]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 泸州老窖</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;5 -->\r\n",
       "<g id=\"edge5\" class=\"edge\"><title>4&#45;&gt;5</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M367.643,-222.907C366.644,-214.558 365.58,-205.671 364.545,-197.02\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"368.012,-196.534 363.349,-187.021 361.062,-197.366 368.012,-196.534\"/>\r\n",
       "</g>\r\n",
       "<!-- 8 -->\r\n",
       "<g id=\"node9\" class=\"node\"><title>8</title>\r\n",
       "<path fill=\"#eb9f68\" stroke=\"black\" d=\"M557.5,-187C557.5,-187 449.5,-187 449.5,-187 443.5,-187 437.5,-181 437.5,-175 437.5,-175 437.5,-116 437.5,-116 437.5,-110 443.5,-104 449.5,-104 449.5,-104 557.5,-104 557.5,-104 563.5,-104 569.5,-110 569.5,-116 569.5,-116 569.5,-175 569.5,-175 569.5,-181 563.5,-187 557.5,-187\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"503.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">脯氨酸 &lt;= 899.564</text>\r\n",
       "<text text-anchor=\"middle\" x=\"503.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.335</text>\r\n",
       "<text text-anchor=\"middle\" x=\"503.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 40</text>\r\n",
       "<text text-anchor=\"middle\" x=\"503.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [32, 6, 2]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"503.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;8 -->\r\n",
       "<g id=\"edge8\" class=\"edge\"><title>4&#45;&gt;8</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M417.95,-222.907C428.416,-213.56 439.637,-203.538 450.396,-193.929\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"453.004,-196.293 458.131,-187.021 448.341,-191.072 453.004,-196.293\"/>\r\n",
       "</g>\r\n",
       "<!-- 6 -->\r\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\r\n",
       "<path fill=\"#8f4fe8\" stroke=\"black\" d=\"M270,-68C270,-68 173,-68 173,-68 167,-68 161,-62 161,-56 161,-56 161,-12 161,-12 161,-6 167,-0 173,-0 173,-0 270,-0 270,-0 276,-0 282,-6 282,-12 282,-12 282,-56 282,-56 282,-62 276,-68 270,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"221.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.18</text>\r\n",
       "<text text-anchor=\"middle\" x=\"221.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 20</text>\r\n",
       "<text text-anchor=\"middle\" x=\"221.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 2, 18]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"221.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 泸州老窖</text>\r\n",
       "</g>\r\n",
       "<!-- 5&#45;&gt;6 -->\r\n",
       "<g id=\"edge6\" class=\"edge\"><title>5&#45;&gt;6</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M307.486,-103.726C295.618,-94.2406 283,-84.1551 271.19,-74.7159\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"273.159,-71.8091 263.163,-68.2996 268.789,-77.2771 273.159,-71.8091\"/>\r\n",
       "</g>\r\n",
       "<!-- 7 -->\r\n",
       "<g id=\"node8\" class=\"node\"><title>7</title>\r\n",
       "<path fill=\"#8139e5\" stroke=\"black\" d=\"M409,-68C409,-68 312,-68 312,-68 306,-68 300,-62 300,-56 300,-56 300,-12 300,-12 300,-6 306,-0 312,-0 312,-0 409,-0 409,-0 415,-0 421,-6 421,-12 421,-12 421,-56 421,-56 421,-62 415,-68 409,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"360.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"360.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 12</text>\r\n",
       "<text text-anchor=\"middle\" x=\"360.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 0, 12]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"360.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 泸州老窖</text>\r\n",
       "</g>\r\n",
       "<!-- 5&#45;&gt;7 -->\r\n",
       "<g id=\"edge7\" class=\"edge\"><title>5&#45;&gt;7</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M359.245,-103.726C359.395,-95.5175 359.553,-86.8595 359.704,-78.56\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"363.208,-78.3619 359.892,-68.2996 356.21,-78.234 363.208,-78.3619\"/>\r\n",
       "</g>\r\n",
       "<!-- 9 -->\r\n",
       "<g id=\"node10\" class=\"node\"><title>9</title>\r\n",
       "<path fill=\"#e6fcef\" stroke=\"black\" d=\"M545,-68C545,-68 456,-68 456,-68 450,-68 444,-62 444,-56 444,-56 444,-12 444,-12 444,-6 450,-0 456,-0 456,-0 545,-0 545,-0 551,-0 557,-6 557,-12 557,-12 557,-56 557,-56 557,-62 551,-68 545,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"500.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.615</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 13</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [5, 6, 2]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 8&#45;&gt;9 -->\r\n",
       "<g id=\"edge9\" class=\"edge\"><title>8&#45;&gt;9</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M502.383,-103.726C502.158,-95.5175 501.921,-86.8595 501.693,-78.56\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"505.185,-78.2 501.412,-68.2996 498.188,-78.3917 505.185,-78.2\"/>\r\n",
       "</g>\r\n",
       "<!-- 10 -->\r\n",
       "<g id=\"node11\" class=\"node\"><title>10</title>\r\n",
       "<path fill=\"#e58139\" stroke=\"black\" d=\"M684,-68C684,-68 587,-68 587,-68 581,-68 575,-62 575,-56 575,-56 575,-12 575,-12 575,-6 581,-0 587,-0 587,-0 684,-0 684,-0 690,-0 696,-6 696,-12 696,-12 696,-56 696,-56 696,-62 690,-68 684,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"635.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"635.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 27</text>\r\n",
       "<text text-anchor=\"middle\" x=\"635.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [27, 0, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"635.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 8&#45;&gt;10 -->\r\n",
       "<g id=\"edge10\" class=\"edge\"><title>8&#45;&gt;10</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M552.652,-103.726C564.087,-94.2406 576.245,-84.1551 587.623,-74.7159\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"589.896,-77.3781 595.358,-68.2996 585.427,-71.9905 589.896,-77.3781\"/>\r\n",
       "</g>\r\n",
       "</g>\r\n",
       "</svg>\r\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x279b6954a20>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph, score = draw_graph(random_state=30,\n",
    "                          splitter='random',\n",
    "                          max_depth=4,\n",
    "                          min_samples_leaf=10\n",
    "                          )\n",
    "graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:52:18.118811Z",
     "start_time": "2019-12-05T00:52:18.100285Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8703703703703703"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:52:19.406905Z",
     "start_time": "2019-12-05T00:52:19.121752Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\r\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\r\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\r\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\r\n",
       " -->\r\n",
       "<!-- Title: Tree Pages: 1 -->\r\n",
       "<svg width=\"635pt\" height=\"552pt\"\r\n",
       " viewBox=\"0.00 0.00 635.00 552.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\r\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 548)\">\r\n",
       "<title>Tree</title>\r\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-548 631,-548 631,4 -4,4\"/>\r\n",
       "<!-- 0 -->\r\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\r\n",
       "<path fill=\"#d8fae6\" stroke=\"black\" d=\"M333,-544C333,-544 220,-544 220,-544 214,-544 208,-538 208,-532 208,-532 208,-473 208,-473 208,-467 214,-461 220,-461 220,-461 333,-461 333,-461 339,-461 345,-467 345,-473 345,-473 345,-532 345,-532 345,-538 339,-544 333,-544\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-528.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">颜色强度 &lt;= 3.82</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-513.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.65</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-498.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 124</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-483.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [37, 54, 33]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"276.5\" y=\"-468.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1 -->\r\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\r\n",
       "<path fill=\"#42e687\" stroke=\"black\" d=\"M278,-425C278,-425 83,-425 83,-425 77,-425 71,-419 71,-413 71,-413 71,-354 71,-354 71,-348 77,-342 83,-342 83,-342 278,-342 278,-342 284,-342 290,-348 290,-354 290,-354 290,-413 290,-413 290,-419 284,-425 278,-425\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">od280/od315稀释葡萄酒 &lt;= 3.36</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.081</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 47</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [2, 45, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"180.5\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;1 -->\r\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M243.193,-460.907C235.747,-451.832 227.779,-442.121 220.106,-432.769\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"222.797,-430.531 213.748,-425.021 217.385,-434.972 222.797,-430.531\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"211.295\" y=\"-446.2\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">True</text>\r\n",
       "</g>\r\n",
       "<!-- 4 -->\r\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\r\n",
       "<path fill=\"#fef9f6\" stroke=\"black\" d=\"M425,-425C425,-425 320,-425 320,-425 314,-425 308,-419 308,-413 308,-413 308,-354 308,-354 308,-348 314,-342 320,-342 320,-342 425,-342 425,-342 431,-342 437,-348 437,-354 437,-354 437,-413 437,-413 437,-419 431,-425 425,-425\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-409.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">类黄酮 &lt;= 1.4</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-394.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.596</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-379.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 77</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-364.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [35, 9, 33]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"372.5\" y=\"-349.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;4 -->\r\n",
       "<g id=\"edge4\" class=\"edge\"><title>0&#45;&gt;4</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M309.807,-460.907C317.253,-451.832 325.221,-442.121 332.894,-432.769\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"335.615,-434.972 339.252,-425.021 330.203,-430.531 335.615,-434.972\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"341.705\" y=\"-446.2\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">False</text>\r\n",
       "</g>\r\n",
       "<!-- 2 -->\r\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\r\n",
       "<path fill=\"#39e581\" stroke=\"black\" d=\"M109,-298.5C109,-298.5 12,-298.5 12,-298.5 6,-298.5 0,-292.5 0,-286.5 0,-286.5 0,-242.5 0,-242.5 0,-236.5 6,-230.5 12,-230.5 12,-230.5 109,-230.5 109,-230.5 115,-230.5 121,-236.5 121,-242.5 121,-242.5 121,-286.5 121,-286.5 121,-292.5 115,-298.5 109,-298.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 42</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 42, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60.5\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;2 -->\r\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M138.867,-341.907C126.87,-330.211 113.79,-317.457 101.813,-305.78\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"104.121,-303.142 94.5174,-298.667 99.234,-308.154 104.121,-303.142\"/>\r\n",
       "</g>\r\n",
       "<!-- 3 -->\r\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\r\n",
       "<path fill=\"#bdf6d5\" stroke=\"black\" d=\"M240,-298.5C240,-298.5 151,-298.5 151,-298.5 145,-298.5 139,-292.5 139,-286.5 139,-286.5 139,-242.5 139,-242.5 139,-236.5 145,-230.5 151,-230.5 151,-230.5 240,-230.5 240,-230.5 246,-230.5 252,-236.5 252,-242.5 252,-242.5 252,-286.5 252,-286.5 252,-292.5 246,-298.5 240,-298.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.48</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 5</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [2, 3, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"195.5\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;3 -->\r\n",
       "<g id=\"edge3\" class=\"edge\"><title>1&#45;&gt;3</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M185.704,-341.907C187.076,-331.204 188.562,-319.615 189.952,-308.776\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"193.448,-309.031 191.248,-298.667 186.505,-308.141 193.448,-309.031\"/>\r\n",
       "</g>\r\n",
       "<!-- 5 -->\r\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\r\n",
       "<path fill=\"#8139e5\" stroke=\"black\" d=\"M407,-298.5C407,-298.5 310,-298.5 310,-298.5 304,-298.5 298,-292.5 298,-286.5 298,-286.5 298,-242.5 298,-242.5 298,-236.5 304,-230.5 310,-230.5 310,-230.5 407,-230.5 407,-230.5 413,-230.5 419,-236.5 419,-242.5 419,-242.5 419,-286.5 419,-286.5 419,-292.5 413,-298.5 407,-298.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-283.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-268.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 33</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-253.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 0, 33]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"358.5\" y=\"-238.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 泸州老窖</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;5 -->\r\n",
       "<g id=\"edge5\" class=\"edge\"><title>4&#45;&gt;5</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M367.643,-341.907C366.362,-331.204 364.975,-319.615 363.678,-308.776\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"367.132,-308.18 362.469,-298.667 360.182,-309.012 367.132,-308.18\"/>\r\n",
       "</g>\r\n",
       "<!-- 6 -->\r\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\r\n",
       "<path fill=\"#eca16c\" stroke=\"black\" d=\"M546,-306C546,-306 449,-306 449,-306 443,-306 437,-300 437,-294 437,-294 437,-235 437,-235 437,-229 443,-223 449,-223 449,-223 546,-223 546,-223 552,-223 558,-229 558,-235 558,-235 558,-294 558,-294 558,-300 552,-306 546,-306\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"497.5\" y=\"-290.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">脯氨酸 &lt;= 737.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"497.5\" y=\"-275.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.325</text>\r\n",
       "<text text-anchor=\"middle\" x=\"497.5\" y=\"-260.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 44</text>\r\n",
       "<text text-anchor=\"middle\" x=\"497.5\" y=\"-245.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [35, 9, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"497.5\" y=\"-230.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;6 -->\r\n",
       "<g id=\"edge6\" class=\"edge\"><title>4&#45;&gt;6</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M415.868,-341.907C425.855,-332.56 436.562,-322.538 446.828,-312.929\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"449.299,-315.41 454.209,-306.021 444.516,-310.299 449.299,-315.41\"/>\r\n",
       "</g>\r\n",
       "<!-- 7 -->\r\n",
       "<g id=\"node8\" class=\"node\"><title>7</title>\r\n",
       "<path fill=\"#4fe88f\" stroke=\"black\" d=\"M475.5,-187C475.5,-187 379.5,-187 379.5,-187 373.5,-187 367.5,-181 367.5,-175 367.5,-175 367.5,-116 367.5,-116 367.5,-110 373.5,-104 379.5,-104 379.5,-104 475.5,-104 475.5,-104 481.5,-104 487.5,-110 487.5,-116 487.5,-116 487.5,-175 487.5,-175 487.5,-181 481.5,-187 475.5,-187\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"427.5\" y=\"-171.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">颜色强度 &lt;= 4.64</text>\r\n",
       "<text text-anchor=\"middle\" x=\"427.5\" y=\"-156.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.18</text>\r\n",
       "<text text-anchor=\"middle\" x=\"427.5\" y=\"-141.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 10</text>\r\n",
       "<text text-anchor=\"middle\" x=\"427.5\" y=\"-126.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [1, 9, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"427.5\" y=\"-111.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 6&#45;&gt;7 -->\r\n",
       "<g id=\"edge7\" class=\"edge\"><title>6&#45;&gt;7</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M473.214,-222.907C467.947,-214.105 462.322,-204.703 456.883,-195.612\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"459.881,-193.805 451.743,-187.021 453.874,-197.399 459.881,-193.805\"/>\r\n",
       "</g>\r\n",
       "<!-- 10 -->\r\n",
       "<g id=\"node11\" class=\"node\"><title>10</title>\r\n",
       "<path fill=\"#e58139\" stroke=\"black\" d=\"M615,-179.5C615,-179.5 518,-179.5 518,-179.5 512,-179.5 506,-173.5 506,-167.5 506,-167.5 506,-123.5 506,-123.5 506,-117.5 512,-111.5 518,-111.5 518,-111.5 615,-111.5 615,-111.5 621,-111.5 627,-117.5 627,-123.5 627,-123.5 627,-167.5 627,-167.5 627,-173.5 621,-179.5 615,-179.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"566.5\" y=\"-164.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"566.5\" y=\"-149.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 34</text>\r\n",
       "<text text-anchor=\"middle\" x=\"566.5\" y=\"-134.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [34, 0, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"566.5\" y=\"-119.3\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 茅台</text>\r\n",
       "</g>\r\n",
       "<!-- 6&#45;&gt;10 -->\r\n",
       "<g id=\"edge10\" class=\"edge\"><title>6&#45;&gt;10</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M521.439,-222.907C528.012,-211.763 535.15,-199.658 541.767,-188.439\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"544.875,-190.059 546.94,-179.667 538.845,-186.503 544.875,-190.059\"/>\r\n",
       "</g>\r\n",
       "<!-- 8 -->\r\n",
       "<g id=\"node9\" class=\"node\"><title>8</title>\r\n",
       "<path fill=\"#6aeca0\" stroke=\"black\" d=\"M407,-68C407,-68 318,-68 318,-68 312,-68 306,-62 306,-56 306,-56 306,-12 306,-12 306,-6 312,-0 318,-0 318,-0 407,-0 407,-0 413,-0 419,-6 419,-12 419,-12 419,-56 419,-56 419,-62 413,-68 407,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"362.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.32</text>\r\n",
       "<text text-anchor=\"middle\" x=\"362.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 5</text>\r\n",
       "<text text-anchor=\"middle\" x=\"362.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [1, 4, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"362.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 7&#45;&gt;8 -->\r\n",
       "<g id=\"edge8\" class=\"edge\"><title>7&#45;&gt;8</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M403.296,-103.726C398.099,-94.9703 392.598,-85.7032 387.375,-76.9051\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"390.381,-75.1121 382.267,-68.2996 384.362,-78.6853 390.381,-75.1121\"/>\r\n",
       "</g>\r\n",
       "<!-- 9 -->\r\n",
       "<g id=\"node10\" class=\"node\"><title>9</title>\r\n",
       "<path fill=\"#39e581\" stroke=\"black\" d=\"M538,-68C538,-68 449,-68 449,-68 443,-68 437,-62 437,-56 437,-56 437,-12 437,-12 437,-6 443,-0 449,-0 449,-0 538,-0 538,-0 544,-0 550,-6 550,-12 550,-12 550,-56 550,-56 550,-62 544,-68 538,-68\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"493.5\" y=\"-52.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">gini = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"493.5\" y=\"-37.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">samples = 5</text>\r\n",
       "<text text-anchor=\"middle\" x=\"493.5\" y=\"-22.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">value = [0, 5, 0]</text>\r\n",
       "<text text-anchor=\"middle\" x=\"493.5\" y=\"-7.8\" font-family=\"Helvetica,sans-Serif\" font-size=\"14.00\">class = 五粮液</text>\r\n",
       "</g>\r\n",
       "<!-- 7&#45;&gt;9 -->\r\n",
       "<g id=\"edge9\" class=\"edge\"><title>7&#45;&gt;9</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M452.076,-103.726C457.354,-94.9703 462.939,-85.7032 468.242,-76.9051\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"471.264,-78.671 473.429,-68.2996 465.269,-75.0574 471.264,-78.671\"/>\r\n",
       "</g>\r\n",
       "</g>\r\n",
       "</svg>\r\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x279b6954ba8>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph, score = draw_graph(random_state=30,\n",
    "                          splitter='best',\n",
    "                          max_depth=4,\n",
    "                          min_samples_leaf=5,\n",
    "                          min_samples_split=2\n",
    "                          )\n",
    "graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:52:45.519112Z",
     "start_time": "2019-12-05T00:52:45.500133Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9259259259259259"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 确定最优参数"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:52:52.085438Z",
     "start_time": "2019-12-05T00:52:51.785570Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "test = []\n",
    "for i in range(10):\n",
    "    clf = tree.DecisionTreeClassifier(max_depth=i+1\n",
    "                                    ,criterion=\"entropy\"\n",
    "                                    ,random_state=30\n",
    "                                    ,splitter=\"random\"\n",
    "                                    )\n",
    "    clf = clf.fit(x_train, y_train)\n",
    "    score = clf.score(x_test, y_test)\n",
    "    test.append(score)\n",
    "plt.plot(range(1,11),test,color=\"red\",label=\"max_depth\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:53:03.251734Z",
     "start_time": "2019-12-05T00:53:03.236347Z"
    }
   },
   "outputs": [],
   "source": [
    "clf = tree.DecisionTreeClassifier(max_depth=4\n",
    "                                ,criterion=\"entropy\"\n",
    "                                ,random_state=30\n",
    "                                ,splitter=\"random\"\n",
    "                                )\n",
    "clf = clf.fit(x_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:53:03.682559Z",
     "start_time": "2019-12-05T00:53:03.669006Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8703703703703703"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.score(x_train, y_train)\n",
    "clf.score(x_test, y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### 重要属性和接口"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "属性是在模型训练之后，能够调用查看的模型的各种性质。对决策树来说，最重要的是feature_importances_，能\n",
    "够查看各个特征对模型的重要性。  \n",
    "\n",
    "sklearn中许多算法的接口都是相似的，比如说我们之前已经用到的fit和score，几乎对每个算法都可以使用。除了\n",
    "这两个接口之外，决策树最常用的接口还有apply和predict。apply中输入测试集返回每个测试样本所在的叶子节点的索引，predict输入测试集返回每个测试样本的标签。返回的内容一目了然并且非常容易。   \n",
    "在这里不得不提的是，**所有接口中要求输入X_train和X_test的部分，输入的特征矩阵必须至少是一个二维矩阵。\n",
    "sklearn不接受任何一维矩阵作为特征矩阵被输入**。如果你的数据的确只有一个特征，那必须用reshape(-1,1)来给\n",
    "矩阵增维；如果你的数据只有一个特征和一个样本，使用reshape(1,-1)来给你的数据增维。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T12:36:37.683089Z",
     "start_time": "2019-11-29T12:36:37.661800Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 3, 16,  3, 19,  5, 14, 20, 16,  3, 14, 16, 10, 20, 20,  6, 20, 10,\n",
       "       20, 20,  6,  5, 17, 16, 14, 20,  6,  5, 20, 14,  3,  9, 20, 10, 20,\n",
       "       16, 14, 19, 16, 10, 20,  3, 20, 17, 20, 20, 20,  3, 20, 14, 20, 20,\n",
       "       16, 16,  3], dtype=int64)"
      ]
     },
     "execution_count": 160,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#apply返回每个测试样本所在的叶子节点的索引\n",
    "clf.apply(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T12:28:42.779070Z",
     "start_time": "2019-11-29T12:28:42.762648Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 1, 2, 1, 2, 1, 0, 1, 2, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 1, 2, 0,\n",
       "       1, 1, 0, 1, 2, 0, 1, 2, 2, 0, 1, 0, 1, 1, 1, 1, 1, 0, 2, 0, 0, 0,\n",
       "       0, 0, 2, 0, 1, 0, 0, 1, 1, 2])"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clf.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T12:31:53.217540Z",
     "start_time": "2019-11-29T12:31:52.925443Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\r\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\r\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\r\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\r\n",
       " -->\r\n",
       "<!-- Title: Tree Pages: 1 -->\r\n",
       "<svg width=\"1096pt\" height=\"477pt\"\r\n",
       " viewBox=\"0.00 0.00 1096.00 477.00\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\r\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 473)\">\r\n",
       "<title>Tree</title>\r\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-473 1092,-473 1092,4 -4,4\"/>\r\n",
       "<!-- 0 -->\r\n",
       "<g id=\"node1\" class=\"node\"><title>0</title>\r\n",
       "<polygon fill=\"#fafefc\" stroke=\"black\" points=\"562.5,-469 437.5,-469 437.5,-401 562.5,-401 562.5,-469\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"500\" y=\"-453.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">类黄酮 &lt;= 1.714</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500\" y=\"-438.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 1.582</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500\" y=\"-423.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 124</text>\r\n",
       "<text text-anchor=\"middle\" x=\"500\" y=\"-408.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [42, 44, 38]</text>\r\n",
       "</g>\r\n",
       "<!-- 1 -->\r\n",
       "<g id=\"node2\" class=\"node\"><title>1</title>\r\n",
       "<polygon fill=\"#b387ef\" stroke=\"black\" points=\"496.5,-365 287.5,-365 287.5,-297 496.5,-297 496.5,-365\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-349.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">od280/od315稀释葡萄酒 &lt;= 2.112</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-334.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.86</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-319.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 53</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-304.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 15, 38]</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;1 -->\r\n",
       "<g id=\"edge1\" class=\"edge\"><title>0&#45;&gt;1</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M464.935,-400.884C455.316,-391.798 444.778,-381.845 434.777,-372.4\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"436.931,-369.621 427.258,-365.299 432.125,-374.71 436.931,-369.621\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"427.972\" y=\"-386.589\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">True</text>\r\n",
       "</g>\r\n",
       "<!-- 12 -->\r\n",
       "<g id=\"node13\" class=\"node\"><title>12</title>\r\n",
       "<polygon fill=\"#f7d8c2\" stroke=\"black\" points=\"667.5,-365 548.5,-365 548.5,-297 667.5,-297 667.5,-365\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-349.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">酒精 &lt;= 12.925</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-334.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.976</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-319.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 71</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-304.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [42, 29, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 0&#45;&gt;12 -->\r\n",
       "<g id=\"edge12\" class=\"edge\"><title>0&#45;&gt;12</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M535.065,-400.884C544.684,-391.798 555.222,-381.845 565.223,-372.4\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"567.875,-374.71 572.742,-365.299 563.069,-369.621 567.875,-374.71\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"572.028\" y=\"-386.589\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">False</text>\r\n",
       "</g>\r\n",
       "<!-- 2 -->\r\n",
       "<g id=\"node3\" class=\"node\"><title>2</title>\r\n",
       "<polygon fill=\"#8c49e7\" stroke=\"black\" points=\"239,-261 127,-261 127,-193 239,-193 239,-261\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-245.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">色调 &lt;= 0.81</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-230.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.391</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-215.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 39</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-200.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 3, 36]</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;2 -->\r\n",
       "<g id=\"edge2\" class=\"edge\"><title>1&#45;&gt;2</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M324.144,-296.884C299.835,-285.02 272.497,-271.678 248.4,-259.918\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"249.859,-256.735 239.337,-255.495 246.789,-263.026 249.859,-256.735\"/>\r\n",
       "</g>\r\n",
       "<!-- 7 -->\r\n",
       "<g id=\"node8\" class=\"node\"><title>7</title>\r\n",
       "<polygon fill=\"#5ae996\" stroke=\"black\" points=\"448,-261 336,-261 336,-193 448,-193 448,-261\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-245.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">酒精 &lt;= 13.198</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-230.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.592</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-215.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 14</text>\r\n",
       "<text text-anchor=\"middle\" x=\"392\" y=\"-200.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 12, 2]</text>\r\n",
       "</g>\r\n",
       "<!-- 1&#45;&gt;7 -->\r\n",
       "<g id=\"edge7\" class=\"edge\"><title>1&#45;&gt;7</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M392,-296.884C392,-288.778 392,-279.982 392,-271.472\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"395.5,-271.299 392,-261.299 388.5,-271.299 395.5,-271.299\"/>\r\n",
       "</g>\r\n",
       "<!-- 3 -->\r\n",
       "<g id=\"node4\" class=\"node\"><title>3</title>\r\n",
       "<polygon fill=\"#8139e5\" stroke=\"black\" points=\"112,-149.5 7.10543e-015,-149.5 7.10543e-015,-96.5 112,-96.5 112,-149.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"56\" y=\"-134.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"56\" y=\"-119.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 31</text>\r\n",
       "<text text-anchor=\"middle\" x=\"56\" y=\"-104.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 0, 31]</text>\r\n",
       "</g>\r\n",
       "<!-- 2&#45;&gt;3 -->\r\n",
       "<g id=\"edge3\" class=\"edge\"><title>2&#45;&gt;3</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M141.767,-192.884C126.978,-181.006 110.344,-167.646 95.6887,-155.876\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"97.7632,-153.053 87.7748,-149.52 93.3799,-158.511 97.7632,-153.053\"/>\r\n",
       "</g>\r\n",
       "<!-- 4 -->\r\n",
       "<g id=\"node5\" class=\"node\"><title>4</title>\r\n",
       "<polygon fill=\"#cdb0f5\" stroke=\"black\" points=\"236,-157 130,-157 130,-89 236,-89 236,-157\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-141.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">类黄酮 &lt;= 0.81</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-126.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.954</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-111.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 8</text>\r\n",
       "<text text-anchor=\"middle\" x=\"183\" y=\"-96.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 3, 5]</text>\r\n",
       "</g>\r\n",
       "<!-- 2&#45;&gt;4 -->\r\n",
       "<g id=\"edge4\" class=\"edge\"><title>2&#45;&gt;4</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M183,-192.884C183,-184.778 183,-175.982 183,-167.472\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"186.5,-167.299 183,-157.299 179.5,-167.299 186.5,-167.299\"/>\r\n",
       "</g>\r\n",
       "<!-- 5 -->\r\n",
       "<g id=\"node6\" class=\"node\"><title>5</title>\r\n",
       "<polygon fill=\"#8139e5\" stroke=\"black\" points=\"112.5,-53 7.5,-53 7.5,-0 112.5,-0 112.5,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"60\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 4</text>\r\n",
       "<text text-anchor=\"middle\" x=\"60\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 0, 4]</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;5 -->\r\n",
       "<g id=\"edge5\" class=\"edge\"><title>4&#45;&gt;5</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M139.981,-88.9485C127.478,-79.3431 113.853,-68.8747 101.449,-59.345\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"103.561,-56.5537 93.4986,-53.2367 99.2961,-62.1046 103.561,-56.5537\"/>\r\n",
       "</g>\r\n",
       "<!-- 6 -->\r\n",
       "<g id=\"node7\" class=\"node\"><title>6</title>\r\n",
       "<polygon fill=\"#7beeab\" stroke=\"black\" points=\"237,-53 131,-53 131,-0 237,-0 237,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"184\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.811</text>\r\n",
       "<text text-anchor=\"middle\" x=\"184\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 4</text>\r\n",
       "<text text-anchor=\"middle\" x=\"184\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 3, 1]</text>\r\n",
       "</g>\r\n",
       "<!-- 4&#45;&gt;6 -->\r\n",
       "<g id=\"edge6\" class=\"edge\"><title>4&#45;&gt;6</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M183.35,-88.9485C183.437,-80.7153 183.531,-71.848 183.619,-63.4814\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"187.122,-63.2732 183.728,-53.2367 180.122,-63.1991 187.122,-63.2732\"/>\r\n",
       "</g>\r\n",
       "<!-- 8 -->\r\n",
       "<g id=\"node9\" class=\"node\"><title>8</title>\r\n",
       "<polygon fill=\"#49e78c\" stroke=\"black\" points=\"366,-157 254,-157 254,-89 366,-89 366,-157\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"310\" y=\"-141.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">类黄酮 &lt;= 0.817</text>\r\n",
       "<text text-anchor=\"middle\" x=\"310\" y=\"-126.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.391</text>\r\n",
       "<text text-anchor=\"middle\" x=\"310\" y=\"-111.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 13</text>\r\n",
       "<text text-anchor=\"middle\" x=\"310\" y=\"-96.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 12, 1]</text>\r\n",
       "</g>\r\n",
       "<!-- 7&#45;&gt;8 -->\r\n",
       "<g id=\"edge8\" class=\"edge\"><title>7&#45;&gt;8</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M365.377,-192.884C358.288,-184.065 350.542,-174.43 343.15,-165.235\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"345.764,-162.9 336.77,-157.299 340.308,-167.286 345.764,-162.9\"/>\r\n",
       "</g>\r\n",
       "<!-- 11 -->\r\n",
       "<g id=\"node12\" class=\"node\"><title>11</title>\r\n",
       "<polygon fill=\"#8139e5\" stroke=\"black\" points=\"489.5,-149.5 384.5,-149.5 384.5,-96.5 489.5,-96.5 489.5,-149.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"437\" y=\"-134.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"437\" y=\"-119.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 1</text>\r\n",
       "<text text-anchor=\"middle\" x=\"437\" y=\"-104.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 0, 1]</text>\r\n",
       "</g>\r\n",
       "<!-- 7&#45;&gt;11 -->\r\n",
       "<g id=\"edge11\" class=\"edge\"><title>7&#45;&gt;11</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M406.61,-192.884C411.414,-181.996 416.766,-169.863 421.625,-158.85\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"424.907,-160.082 425.741,-149.52 418.503,-157.256 424.907,-160.082\"/>\r\n",
       "</g>\r\n",
       "<!-- 9 -->\r\n",
       "<g id=\"node10\" class=\"node\"><title>9</title>\r\n",
       "<polygon fill=\"#8139e5\" stroke=\"black\" points=\"361.5,-53 256.5,-53 256.5,-0 361.5,-0 361.5,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"309\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"309\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 1</text>\r\n",
       "<text text-anchor=\"middle\" x=\"309\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 0, 1]</text>\r\n",
       "</g>\r\n",
       "<!-- 8&#45;&gt;9 -->\r\n",
       "<g id=\"edge9\" class=\"edge\"><title>8&#45;&gt;9</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M309.65,-88.9485C309.563,-80.7153 309.469,-71.848 309.381,-63.4814\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"312.878,-63.1991 309.272,-53.2367 305.878,-63.2732 312.878,-63.1991\"/>\r\n",
       "</g>\r\n",
       "<!-- 10 -->\r\n",
       "<g id=\"node11\" class=\"node\"><title>10</title>\r\n",
       "<polygon fill=\"#39e581\" stroke=\"black\" points=\"492,-53 380,-53 380,-0 492,-0 492,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"436\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"436\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 12</text>\r\n",
       "<text text-anchor=\"middle\" x=\"436\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 12, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 8&#45;&gt;10 -->\r\n",
       "<g id=\"edge10\" class=\"edge\"><title>8&#45;&gt;10</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M354.069,-88.9485C366.876,-79.3431 380.834,-68.8747 393.54,-59.345\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"395.784,-62.0368 401.684,-53.2367 391.584,-56.4367 395.784,-62.0368\"/>\r\n",
       "</g>\r\n",
       "<!-- 13 -->\r\n",
       "<g id=\"node14\" class=\"node\"><title>13</title>\r\n",
       "<polygon fill=\"#41e686\" stroke=\"black\" points=\"664,-261 552,-261 552,-193 664,-193 664,-261\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-245.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">灰 &lt;= 2.26</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-230.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.235</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-215.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 26</text>\r\n",
       "<text text-anchor=\"middle\" x=\"608\" y=\"-200.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [1, 25, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 12&#45;&gt;13 -->\r\n",
       "<g id=\"edge13\" class=\"edge\"><title>12&#45;&gt;13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M608,-296.884C608,-288.778 608,-279.982 608,-271.472\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"611.5,-271.299 608,-261.299 604.5,-271.299 611.5,-271.299\"/>\r\n",
       "</g>\r\n",
       "<!-- 18 -->\r\n",
       "<g id=\"node19\" class=\"node\"><title>18</title>\r\n",
       "<polygon fill=\"#e88d4c\" stroke=\"black\" points=\"961,-261 849,-261 849,-193 961,-193 961,-261\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-245.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">镁 &lt;= 89.942</text>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-230.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.433</text>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-215.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 45</text>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-200.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [41, 4, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 12&#45;&gt;18 -->\r\n",
       "<g id=\"edge18\" class=\"edge\"><title>12&#45;&gt;18</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M667.745,-309.481C717.396,-292.43 787.954,-268.197 839.21,-250.594\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"840.548,-253.836 848.869,-247.277 838.275,-247.215 840.548,-253.836\"/>\r\n",
       "</g>\r\n",
       "<!-- 14 -->\r\n",
       "<g id=\"node15\" class=\"node\"><title>14</title>\r\n",
       "<polygon fill=\"#39e581\" stroke=\"black\" points=\"620,-149.5 508,-149.5 508,-96.5 620,-96.5 620,-149.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"564\" y=\"-134.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"564\" y=\"-119.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 14</text>\r\n",
       "<text text-anchor=\"middle\" x=\"564\" y=\"-104.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 14, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 13&#45;&gt;14 -->\r\n",
       "<g id=\"edge14\" class=\"edge\"><title>13&#45;&gt;14</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M593.714,-192.884C589.018,-181.996 583.784,-169.863 579.033,-158.85\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"582.183,-157.316 575.009,-149.52 575.756,-160.088 582.183,-157.316\"/>\r\n",
       "</g>\r\n",
       "<!-- 15 -->\r\n",
       "<g id=\"node16\" class=\"node\"><title>15</title>\r\n",
       "<polygon fill=\"#4be78c\" stroke=\"black\" points=\"834,-157 638,-157 638,-89 834,-89 834,-157\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"736\" y=\"-141.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">od280/od315稀释葡萄酒 &lt;= 3.4</text>\r\n",
       "<text text-anchor=\"middle\" x=\"736\" y=\"-126.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.414</text>\r\n",
       "<text text-anchor=\"middle\" x=\"736\" y=\"-111.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 12</text>\r\n",
       "<text text-anchor=\"middle\" x=\"736\" y=\"-96.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [1, 11, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 13&#45;&gt;15 -->\r\n",
       "<g id=\"edge15\" class=\"edge\"><title>13&#45;&gt;15</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M649.558,-192.884C661.295,-183.531 674.184,-173.259 686.346,-163.568\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"688.573,-166.269 694.213,-157.299 684.211,-160.794 688.573,-166.269\"/>\r\n",
       "</g>\r\n",
       "<!-- 16 -->\r\n",
       "<g id=\"node17\" class=\"node\"><title>16</title>\r\n",
       "<polygon fill=\"#39e581\" stroke=\"black\" points=\"728,-53 616,-53 616,-0 728,-0 728,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"672\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"672\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 11</text>\r\n",
       "<text text-anchor=\"middle\" x=\"672\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 11, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 15&#45;&gt;16 -->\r\n",
       "<g id=\"edge16\" class=\"edge\"><title>15&#45;&gt;16</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M713.616,-88.9485C707.606,-80.0749 701.098,-70.4648 695.053,-61.5388\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"697.936,-59.5539 689.43,-53.2367 692.14,-63.4792 697.936,-59.5539\"/>\r\n",
       "</g>\r\n",
       "<!-- 17 -->\r\n",
       "<g id=\"node18\" class=\"node\"><title>17</title>\r\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"851.5,-53 746.5,-53 746.5,-0 851.5,-0 851.5,-53\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"799\" y=\"-37.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"799\" y=\"-22.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 1</text>\r\n",
       "<text text-anchor=\"middle\" x=\"799\" y=\"-7.8\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [1, 0, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 15&#45;&gt;17 -->\r\n",
       "<g id=\"edge17\" class=\"edge\"><title>15&#45;&gt;17</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M758.034,-88.9485C763.889,-80.1664 770.225,-70.6629 776.123,-61.815\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"779.207,-63.4987 781.842,-53.2367 773.383,-59.6158 779.207,-63.4987\"/>\r\n",
       "</g>\r\n",
       "<!-- 19 -->\r\n",
       "<g id=\"node20\" class=\"node\"><title>19</title>\r\n",
       "<polygon fill=\"#39e581\" stroke=\"black\" points=\"957.5,-149.5 852.5,-149.5 852.5,-96.5 957.5,-96.5 957.5,-149.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-134.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-119.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 4</text>\r\n",
       "<text text-anchor=\"middle\" x=\"905\" y=\"-104.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [0, 4, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 18&#45;&gt;19 -->\r\n",
       "<g id=\"edge19\" class=\"edge\"><title>18&#45;&gt;19</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M905,-192.884C905,-182.326 905,-170.597 905,-159.854\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"908.5,-159.52 905,-149.52 901.5,-159.52 908.5,-159.52\"/>\r\n",
       "</g>\r\n",
       "<!-- 20 -->\r\n",
       "<g id=\"node21\" class=\"node\"><title>20</title>\r\n",
       "<polygon fill=\"#e58139\" stroke=\"black\" points=\"1088,-149.5 976,-149.5 976,-96.5 1088,-96.5 1088,-149.5\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"1032\" y=\"-134.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">entropy = 0.0</text>\r\n",
       "<text text-anchor=\"middle\" x=\"1032\" y=\"-119.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">samples = 41</text>\r\n",
       "<text text-anchor=\"middle\" x=\"1032\" y=\"-104.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">value = [41, 0, 0]</text>\r\n",
       "</g>\r\n",
       "<!-- 18&#45;&gt;20 -->\r\n",
       "<g id=\"edge20\" class=\"edge\"><title>18&#45;&gt;20</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M946.233,-192.884C961.022,-181.006 977.656,-167.646 992.311,-155.876\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"994.62,-158.511 1000.23,-149.52 990.237,-153.053 994.62,-158.511\"/>\r\n",
       "</g>\r\n",
       "</g>\r\n",
       "</svg>\r\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x20bc2b58240>"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dot_data = tree.export_graphviz(clf,\n",
    "                                feature_names=feature_names,\n",
    "                                filled=True)\n",
    "graph = graphviz.Source(dot_data)\n",
    "graph"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T12:33:29.540850Z",
     "start_time": "2019-11-29T12:33:29.530434Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'images/graph.gv'"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph.save('images/graph.gv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-11-29T12:34:00.785466Z",
     "start_time": "2019-11-29T12:34:00.344736Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'images\\\\graph.gv.pdf'"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "graph.view()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# DecisionTreeRegressor(线性回归决策树)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:56:43.524348Z",
     "start_time": "2019-12-05T00:56:43.507430Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.tree import DecisionTreeRegressor\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 重要属性接口\n",
    "```python\n",
    "Init signature:\n",
    "DecisionTreeRegressor(\n",
    "    criterion='mse',\n",
    "    splitter='best',\n",
    "    max_depth=None,\n",
    "    min_samples_split=2,\n",
    "    min_samples_leaf=1,\n",
    "    min_weight_fraction_leaf=0.0,\n",
    "    max_features=None,\n",
    "    random_state=None,\n",
    "    max_leaf_nodes=None,\n",
    "    min_impurity_decrease=0.0,\n",
    "    min_impurity_split=None,\n",
    "    presort=False,\n",
    ")\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T00:58:46.956183Z",
     "start_time": "2019-12-05T00:58:46.936673Z"
    }
   },
   "source": [
    "### criterion(评估标准)\n",
    "1. 输入\"mse\"使用均方误差mean squared error(MSE)，父节点和叶子节点之间的均方误差的差额将被用来作为\n",
    "特征选择的标准，这种方法通过使用叶子节点的均值来最小化L2损失\n",
    "2. 输入“friedman_mse”使用费尔德曼均方误差，这种指标使用弗里德曼针对潜在分枝中的问题改进后的均方误差\n",
    "3. 输入\"mae\"使用绝对平均误差MAE（mean absolute error），这种指标使用叶节点的中值来最小化L1损失\n",
    "属性中最重要的依然是feature_importances_，接口依然是apply, fit, predict, score最核心"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$$MSE = \\dfrac{1}{N}\\displaystyle \\sum_{i=1}^{N}(f_i - y_i)^2$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "其中N是样本数量，i是每一个数据样本，fi是模型回归出的数值，yi是样本点i实际的数值标签。所以MSE的本质，\n",
    "其实是样本真实数据与回归结果的差异（类似于损失函数）。在回归树中，MSE不只是我们的分枝质量衡量指标，也是我们最常用的衡\n",
    "量回归树回归质量的指标，当我们在使用交叉验证，或者其他方式获取回归树的结果时，我们往往选择均方误差作\n",
    "为我们的评估（在分类树中这个指标是score代表的预测准确率）。在回归中，我们追求的是，**MSE越小越好**。\n",
    "然而，**回归树的接口score返回的是R平方，并不是MSE**。R平方被定义如下:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "$$R^2 = 1 - \\dfrac{u}{v} \\; \\;\\;其中, \\\\\n",
    "u = \\displaystyle \\sum_{i=1}^{N}(f_i - y_i)^2\\;\\;\\;\\;v = \\sum_{i=1}^{N}(y_i - \\hat{y})^2$$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "其中u为残差平方和(MSE * N), v是总平方和，N是样本数量，i是每一个数据样本，fi是模型回归出的数值，yi\n",
    "是样本点i实际的数值标签。$\\hat{y}$是**真实数值标签的平均数**。R平方可以为正为负（如果模型的残差平方和远远大于\n",
    "模型的总平方和，模型非常糟糕，R平方就会为负），而均方误差永远为正\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "值得一提的是，**虽然均方误差永远为正，但是sklearn当中使用均方误差作为评判标准时，却是计算”负均方误\n",
    "差“（neg_mean_squared_error）**。这是因为sklearn在计算模型评估指标的时候，会考虑指标本身的性质，均\n",
    "方误差本身是一种误差，所以被sklearn划分为模型的一种损失(loss)，因此在sklearn当中，都以负数表示。真正的\n",
    "均方误差MSE的数值，其实就是neg_mean_squared_error去掉负号的数字\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 交叉验证\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">交叉验证是用来观察模型的稳定性的一种方法，我们将数据划分为n份，依次使用其中一份作为测试集，其他n-1份\n",
    "作为训练集，多次计算模型的精确性来评估模型的平均准确程度。训练集和测试集的划分会干扰模型的结果，因此\n",
    "用交叉验证n次的结果求出的平均值，是对模型效果的一个更好的度量"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 使用sklearn实现交叉验证"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T01:26:42.427592Z",
     "start_time": "2019-12-05T01:26:42.418699Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.datasets import load_boston\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.tree import DecisionTreeRegressor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T01:31:50.452173Z",
     "start_time": "2019-12-05T01:31:50.328168Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([ 0.52939335,  0.60461936, -1.60907519,  0.4356399 ,  0.77280671,\n",
       "        0.40597035,  0.23656049,  0.38709149, -2.06488186, -0.95162992])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "boston = load_boston()\n",
    "regressor = DecisionTreeRegressor(random_state=0)\n",
    "cross_val_score(regressor, boston.data, boston.target, cv=10)\n",
    "                #, scoring='neg_mean_squared_error')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "sklearn fit只支持二维数据\n",
    "reval降维"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 示例：一维回归的图像绘制"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:20:58.453186Z",
     "start_time": "2019-12-05T02:20:58.443291Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:21:28.652000Z",
     "start_time": "2019-12-05T02:21:28.645260Z"
    }
   },
   "outputs": [],
   "source": [
    "# 设置随机种子\n",
    "rng = np.random.RandomState(1)\n",
    "# 生成0-5区间内的随机数据 with shape (80, 1)\n",
    "x = np.sort(5 * rng.rand(80,1), axis=0)\n",
    "# 生成x对应的sin(y)\n",
    "y = np.sin(x).ravel()  # ravel() 将数据降到一维"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:22:43.313524Z",
     "start_time": "2019-12-05T02:22:43.295533Z"
    }
   },
   "outputs": [],
   "source": [
    "# 对y添加噪音\n",
    "y[::5] += 3 * (0.5 - rng.rand(16))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:26:25.206445Z",
     "start_time": "2019-12-05T02:26:24.917459Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(dpi=150, figsize=(6,3))\n",
    "plt.scatter(x, y, s=20, edgecolors='black', c='darkorange', label='data')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:28:16.761405Z",
     "start_time": "2019-12-05T02:28:16.754493Z"
    }
   },
   "outputs": [],
   "source": [
    "# 进行线性回归决策树拟合，创建3个max_path不同的决策树\n",
    "regressor_1 = DecisionTreeRegressor(max_depth=2)\n",
    "regressor_2 = DecisionTreeRegressor(max_depth=3)\n",
    "regressor_3 = DecisionTreeRegressor(max_depth=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:30:10.069639Z",
     "start_time": "2019-12-05T02:30:10.051158Z"
    }
   },
   "outputs": [],
   "source": [
    "reg_1 = regressor_1.fit(x, y)\n",
    "reg_2 = regressor_2.fit(x, y)\n",
    "reg_3 = regressor_3.fit(x, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:31:08.036909Z",
     "start_time": "2019-12-05T02:31:08.026988Z"
    }
   },
   "outputs": [],
   "source": [
    "# 创建测试集\n",
    "x_test = np.linspace(0, 5, 100)[:, np.newaxis]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:31:53.084047Z",
     "start_time": "2019-12-05T02:31:53.069073Z"
    }
   },
   "outputs": [],
   "source": [
    "# 拟合数据\n",
    "y_1 = reg_1.predict(x_test)\n",
    "y_2 = reg_2.predict(x_test)\n",
    "y_3 = reg_3.predict(x_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T02:36:01.991241Z",
     "start_time": "2019-12-05T02:36:01.588982Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 900x450 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 绘图\n",
    "fig = plt.figure(dpi=150, figsize=(6,3))\n",
    "plt.scatter(x, y, s=20, edgecolors='black', c='darkorange', label='data')\n",
    "# plt.plot(x_test, y_1, color='cornflowerblue', label='max_depth=2', linewidth=2)\n",
    "plt.plot(x_test, y_2, color='green', label='max_depth=3', linewidth=2)\n",
    "plt.plot(x_test, y_3,color='red', label='max_depth=5', linewidth=2)\n",
    "\n",
    "plt.xlabel('x')\n",
    "plt.ylabel('y')\n",
    "plt.title('difference of max depth')\n",
    "plt.legend(loc='best')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 示例：泰坦尼克幸存者的预测"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:04:31.497728Z",
     "start_time": "2019-12-05T03:04:31.490929Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 加载数据"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:06:26.843216Z",
     "start_time": "2019-12-05T03:06:26.825812Z"
    }
   },
   "outputs": [],
   "source": [
    "data = pd.read_csv('data/titanic_data.csv', index_col=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:03:56.026649Z",
     "start_time": "2019-12-05T03:03:56.000868Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Name</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Ticket</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Cabin</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PassengerId</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Braund, Mr. Owen Harris</td>\n",
       "      <td>male</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>A/5 21171</td>\n",
       "      <td>7.2500</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Cumings, Mrs. John Bradley (Florence Briggs Th...</td>\n",
       "      <td>female</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>PC 17599</td>\n",
       "      <td>71.2833</td>\n",
       "      <td>C85</td>\n",
       "      <td>C</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>Heikkinen, Miss. Laina</td>\n",
       "      <td>female</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>STON/O2. 3101282</td>\n",
       "      <td>7.9250</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>Futrelle, Mrs. Jacques Heath (Lily May Peel)</td>\n",
       "      <td>female</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>113803</td>\n",
       "      <td>53.1000</td>\n",
       "      <td>C123</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>Allen, Mr. William Henry</td>\n",
       "      <td>male</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>373450</td>\n",
       "      <td>8.0500</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Survived  Pclass  \\\n",
       "PassengerId                     \n",
       "1                   0       3   \n",
       "2                   1       1   \n",
       "3                   1       3   \n",
       "4                   1       1   \n",
       "5                   0       3   \n",
       "\n",
       "                                                          Name     Sex   Age  \\\n",
       "PassengerId                                                                    \n",
       "1                                      Braund, Mr. Owen Harris    male  22.0   \n",
       "2            Cumings, Mrs. John Bradley (Florence Briggs Th...  female  38.0   \n",
       "3                                       Heikkinen, Miss. Laina  female  26.0   \n",
       "4                 Futrelle, Mrs. Jacques Heath (Lily May Peel)  female  35.0   \n",
       "5                                     Allen, Mr. William Henry    male  35.0   \n",
       "\n",
       "             SibSp  Parch            Ticket     Fare Cabin Embarked  \n",
       "PassengerId                                                          \n",
       "1                1      0         A/5 21171   7.2500   NaN        S  \n",
       "2                1      0          PC 17599  71.2833   C85        C  \n",
       "3                0      0  STON/O2. 3101282   7.9250   NaN        S  \n",
       "4                1      0            113803  53.1000  C123        S  \n",
       "5                0      0            373450   8.0500   NaN        S  "
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:03:56.406010Z",
     "start_time": "2019-12-05T03:03:56.390136Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 891 entries, 1 to 891\n",
      "Data columns (total 11 columns):\n",
      "Survived    891 non-null int64\n",
      "Pclass      891 non-null int64\n",
      "Name        891 non-null object\n",
      "Sex         891 non-null object\n",
      "Age         714 non-null float64\n",
      "SibSp       891 non-null int64\n",
      "Parch       891 non-null int64\n",
      "Ticket      891 non-null object\n",
      "Fare        891 non-null float64\n",
      "Cabin       204 non-null object\n",
      "Embarked    889 non-null object\n",
      "dtypes: float64(2), int64(4), object(5)\n",
      "memory usage: 83.5+ KB\n"
     ]
    }
   ],
   "source": [
    "data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 数据清洗"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:06:29.675188Z",
     "start_time": "2019-12-05T03:06:29.667189Z"
    }
   },
   "outputs": [],
   "source": [
    "# 删除无效数据\n",
    "data = data.drop(['Ticket', 'Cabin', 'Name'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:07:15.581251Z",
     "start_time": "2019-12-05T03:07:15.575519Z"
    }
   },
   "outputs": [],
   "source": [
    "# 填充数据\n",
    "data['Age'] = data['Age'].fillna(data['Age'].mean())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:08:31.156844Z",
     "start_time": "2019-12-05T03:08:31.098402Z"
    }
   },
   "outputs": [],
   "source": [
    "# 对于有些缺失值的数据，直接drop\n",
    "data = data.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:15:21.558098Z",
     "start_time": "2019-12-05T03:15:21.190069Z"
    }
   },
   "outputs": [],
   "source": [
    "# 将分类数据转换成数值数据\n",
    "data.head()\n",
    "# 二分类变量转换成数据值型数据\n",
    "data['Sex'] = (data[\"Sex\"]== \"male\").astype(\"int\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:20:47.428975Z",
     "start_time": "2019-12-05T03:20:47.419100Z"
    }
   },
   "outputs": [],
   "source": [
    "# 多分类变量转换成数值型数据\n",
    "labels = data['Embarked'].unique().tolist()\n",
    "data['Embarked'] =  data['Embarked'].apply(lambda x: labels.index(x))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:20:58.511256Z",
     "start_time": "2019-12-05T03:20:58.482413Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Survived</th>\n",
       "      <th>Pclass</th>\n",
       "      <th>Sex</th>\n",
       "      <th>Age</th>\n",
       "      <th>SibSp</th>\n",
       "      <th>Parch</th>\n",
       "      <th>Fare</th>\n",
       "      <th>Embarked</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>PassengerId</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>22.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>7.2500</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>38.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>71.2833</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>26.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7.9250</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>53.1000</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>35.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8.0500</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             Survived  Pclass  Sex   Age  SibSp  Parch     Fare  Embarked\n",
       "PassengerId                                                              \n",
       "1                   0       3    1  22.0      1      0   7.2500         0\n",
       "2                   1       1    0  38.0      1      0  71.2833         1\n",
       "3                   1       3    0  26.0      0      0   7.9250         0\n",
       "4                   1       1    0  35.0      1      0  53.1000         0\n",
       "5                   0       3    1  35.0      0      0   8.0500         0"
      ]
     },
     "execution_count": 114,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 提取标签和特征矩阵，划分测试集和训练集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:23:31.163624Z",
     "start_time": "2019-12-05T03:23:31.155059Z"
    }
   },
   "outputs": [],
   "source": [
    "x = data.iloc[:, data.columns != 'Survived']\n",
    "y = data.iloc[:, data.columns == 'Survived']\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:24:47.708086Z",
     "start_time": "2019-12-05T03:24:47.692986Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "Xtrain, Xtest, ytrain, ytest = train_test_split(x, y, test_size=.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:26:51.039760Z",
     "start_time": "2019-12-05T03:26:51.032322Z"
    }
   },
   "outputs": [],
   "source": [
    "# 修正索引\n",
    "for i in [Xtrain, Xtest, ytrain, ytest]:\n",
    "    i.index = range(i.shape[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 导入模型，粗略地跑一下"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:28:38.462708Z",
     "start_time": "2019-12-05T03:28:38.456461Z"
    }
   },
   "outputs": [],
   "source": [
    "clf = DecisionTreeClassifier(random_state=25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:28:50.592603Z",
     "start_time": "2019-12-05T03:28:50.531087Z"
    }
   },
   "outputs": [],
   "source": [
    "clf = clf.fit(Xtrain, ytrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:29:34.664415Z",
     "start_time": "2019-12-05T03:29:34.648740Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7602996254681648"
      ]
     },
     "execution_count": 129,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_ = clf.score(Xtest, ytest)\n",
    "score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:30:28.864175Z",
     "start_time": "2019-12-05T03:30:28.742594Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7491452991452991"
      ]
     },
     "execution_count": 130,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_ = cross_val_score(clf, Xtest, ytest, cv=10).mean()\n",
    "score_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:38:53.309435Z",
     "start_time": "2019-12-05T03:38:53.304971Z"
    }
   },
   "outputs": [],
   "source": [
    "# import graphviz\n",
    "# from sklearn import tree\n",
    "# dot_data = tree.export_graphviz(clf, filled=True, rounded=True, feature_names=x.columns, class_names=['Survived', 'death'])\n",
    "# graphviz.Source(dot_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:37:06.741406Z",
     "start_time": "2019-12-05T03:37:06.732506Z"
    }
   },
   "source": [
    "## 在不同max_depth下观察模型拟合状况"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:50:34.966873Z",
     "start_time": "2019-12-05T03:50:33.652404Z"
    }
   },
   "outputs": [],
   "source": [
    "# 创建一个列表用于接收拟合后的分数\n",
    "tr = []\n",
    "# te用于接收交叉验证后的分数\n",
    "te = []\n",
    "\n",
    "for i in range(10):\n",
    "    # 这里为什么使用“entropy”？因为我们注意到，在最大深度=3的时候，模型拟合不足，在训练集和测试集上的表现接\n",
    "    # 近，但却都不是非常理想，只能够达到83%左右，所以我们要使用entropy。\n",
    "    temp_clf = DecisionTreeClassifier(random_state=25,\n",
    "                                 max_depth=i+1,\n",
    "                                criterion='entropy')\n",
    "    temp_clf = temp_clf.fit(Xtrain, ytrain)\n",
    "    \n",
    "    temp_score = temp_clf.score(Xtrain, ytrain)\n",
    "    temp_score2 = cross_val_score(temp_clf, x, y, cv=10).mean()\n",
    "    \n",
    "    tr.append(temp_score)\n",
    "    te.append(temp_score2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T03:50:37.600957Z",
     "start_time": "2019-12-05T03:50:37.594250Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9115755627009646\n",
      "0.8177860061287026\n"
     ]
    }
   ],
   "source": [
    "print(max(tr))\n",
    "print(max(te))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T06:58:49.681619Z",
     "start_time": "2019-12-05T06:58:49.101263Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 绘图\n",
    "plt.plot(range(1,11), tr, color='red', label='performance in train')\n",
    "plt.plot(range(1,11), te, color='blue', label='performance in test')\n",
    "plt.legend()\n",
    "plt.xticks(range(1, 11))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 使用网格搜索调整参数\n",
    ">网格搜索类似于我们调参数，只是参数可以一次性调很多个，而且全部可以交给程序去跑。  \n",
    "\n",
    "网格搜索的优缺点很明显：可以自动调参，但是，当我们把参数传进去后，网格搜索就一定会按照传进去的参数进行选择，不会删除某些参数。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:16:35.826069Z",
     "start_time": "2019-12-05T07:16:35.817668Z"
    }
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:46:27.540604Z",
     "start_time": "2019-12-05T07:28:22.318206Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "D:\\anaconda\\lib\\site-packages\\sklearn\\model_selection\\_search.py:813: DeprecationWarning: The default of the `iid` parameter will change from True to False in version 0.22 and will be removed in 0.24. This will change numeric results when test-set sizes are unequal.\n",
      "  DeprecationWarning)\n"
     ]
    }
   ],
   "source": [
    "# gini的取值范围在0 - 0.5\n",
    "# entropy的取值范围在0 - 1\n",
    "gini_thresholds = np.linspace(0, 0.3, 20)\n",
    "parameters = {\n",
    "    'max_depth':[*range(1, 10), None],\n",
    "    'criterion':['gini', 'entropy'],\n",
    "    'splitter':['best', 'random'],\n",
    "    'min_samples_leaf':[*range(1,8)],\n",
    "    'min_samples_split':[*range(2,6)],\n",
    "    'min_impurity_decrease':gini_thresholds,\n",
    "}\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "clf = DecisionTreeClassifier(random_state=25)\n",
    "GS = GridSearchCV(clf, parameters, cv=5)  # 模型， 参数字典， 交叉验证划分\n",
    "GS = GS.fit(Xtrain, ytrain)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:50:53.646078Z",
     "start_time": "2019-12-05T07:50:53.638132Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'criterion': 'entropy', 'max_depth': 7, 'min_impurity_decrease': 0.0, 'min_samples_leaf': 7, 'min_samples_split': 2, 'splitter': 'best'}\n",
      "0.8247588424437299\n"
     ]
    }
   ],
   "source": [
    "# 输出最佳的参数组合\n",
    "print(GS.best_params_)\n",
    "\n",
    "# 输出最好的分数\n",
    "print(GS.best_score_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 决策树的优缺点"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 优点"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. 易于理解和解释，因为树木可以画出来被看见\n",
    "2. 需要很少的数据准备。其他很多算法通常都需要数据规范化，需要创建虚拟变量并删除空值等。但请注意，\n",
    "**sklearn中的决策树模块不支持对缺失值的处理**。\n",
    "3. 使用树的成本（比如说，在预测数据的时候）是用于训练树的数据点的数量的对数，相比于其他算法，这是\n",
    "一个很低的成本。\n",
    "4. 能够同时处理数字和分类数据，既可以做回归又可以做分类。其他技术通常专门用于分析仅具有一种变量类\n",
    "型的数据集。\n",
    "5. 能够处理多输出问题，即含有多个标签的问题，注意与一个标签中含有多种标签分类的问题区别开\n",
    "6. 是一个白盒模型，结果很容易能够被解释。如果在模型中可以观察到给定的情况，则可以通过布尔逻辑轻松\n",
    "解释条件。相反，在黑盒模型中（例如，在人工神经网络中），结果可能更难以解释。\n",
    "7. 可以使用统计测试验证模型，这让我们可以考虑模型的可靠性。\n",
    "8. 即使其假设在某种程度上违反了生成数据的真实模型，也能够表现良好"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 缺点"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1. 决策树学习者可能创建过于复杂的树，这些树不能很好地推广数据。这称为过度拟合。修剪，设置叶节点所\n",
    "需的最小样本数或设置树的最大深度等机制是避免此问题所必需的，而这些参数的整合和调整对初学者来说\n",
    "会比较晦涩\n",
    "2. 决策树可能不稳定，数据中微小的变化可能导致生成完全不同的树，这个问题需要通过集成算法来解决。\n",
    "3. 决策树的学习是基于贪婪算法，它靠优化局部最优（每个节点的最优）来试图达到整体的最优，但这种做法\n",
    "不能保证返回全局最优决策树。这个问题也可以由集成算法来解决，在随机森林中，特征和样本会在分枝过\n",
    "程中被随机采样。\n",
    "4. 有些概念很难学习，因为决策树不容易表达它们，例如XOR，奇偶校验或多路复用器问题。\n",
    "5. 如果标签中的某些类占主导地位，决策树学习者会创建偏向主导类的树。因此，建议在拟合决策树之前平衡\n",
    "数据集。(例如，ID)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:46:28.426487Z",
     "start_time": "2019-12-05T07:46:28.418594Z"
    }
   },
   "source": [
    "# 实例：分类树在合成数集上的表现"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:53:52.946282Z",
     "start_time": "2019-12-05T07:53:52.889738Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.colors import ListedColormap\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.datasets import make_moons, make_circles, make_classification\n",
    "from sklearn.tree import DecisionTreeClassifier"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 生成三种数据集"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:56:59.674947Z",
     "start_time": "2019-12-05T07:56:59.666508Z"
    }
   },
   "outputs": [],
   "source": [
    "# make_classification 生成二分类数据集\n",
    "X,y = make_classification(n_samples=100, # 生成100个样本\n",
    "                         n_features=2, # 生成2个特征\n",
    "                         n_redundant=0,  # 设置0个冗余特征\n",
    "                         n_informative=2,  # 设置含主要信息的特征为2\n",
    "                         random_state=1, # 设置随机种子,\n",
    "                         n_clusters_per_class=1  # 每个簇包含的标签类别有1个\n",
    "                         )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 191,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T07:59:12.987290Z",
     "start_time": "2019-12-05T07:59:12.943128Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([[-1.04948638,  0.8786438 ],\n",
       "        [ 0.8780991 ,  0.89551051],\n",
       "        [ 0.95928819,  1.03967316],\n",
       "        [-1.18013412,  1.12062155],\n",
       "        [-0.90731836,  1.06040861],\n",
       "        [-0.62754626,  1.37661405],\n",
       "        [ 0.82178321,  1.18947778],\n",
       "        [-1.20376927,  0.78801845],\n",
       "        [ 0.77151441,  0.345716  ],\n",
       "        [-0.493857  ,  1.66752297],\n",
       "        [ 1.17456965,  0.69211449],\n",
       "        [ 0.82798737,  0.84806927],\n",
       "        [ 0.89189141,  0.87699465],\n",
       "        [ 0.90394545,  0.35413361],\n",
       "        [ 0.94459507,  1.12641981],\n",
       "        [ 0.9317172 ,  0.78344054],\n",
       "        [-0.96853969,  1.27991386],\n",
       "        [-1.09230828,  0.92686981],\n",
       "        [-0.68243019,  1.4468218 ],\n",
       "        [-0.76045809,  0.61363671],\n",
       "        [-1.1146902 ,  1.13545112],\n",
       "        [ 1.32970299,  1.58463774],\n",
       "        [-1.39003042,  0.85771953],\n",
       "        [ 1.01313574,  1.16777676],\n",
       "        [ 1.22382712,  1.98519313],\n",
       "        [-1.20257258,  0.69174373],\n",
       "        [ 0.96435056,  1.77547698],\n",
       "        [ 1.02600767,  0.87226739],\n",
       "        [ 1.10863359,  1.16755486],\n",
       "        [-0.84967187,  0.88838944],\n",
       "        [-1.03048932,  0.61839045],\n",
       "        [-1.278618  ,  0.98379959],\n",
       "        [-0.98437725,  0.93621125],\n",
       "        [ 0.91588404,  0.79032389],\n",
       "        [ 1.13872896,  0.85729649],\n",
       "        [-1.01366961,  0.47330777],\n",
       "        [ 0.84917607,  1.36586461],\n",
       "        [ 1.03436581,  2.07582971],\n",
       "        [ 1.24328724,  1.07258847],\n",
       "        [ 0.84146721,  0.89891716],\n",
       "        [-1.47533915,  0.86146256],\n",
       "        [-1.23912256,  0.94781411],\n",
       "        [ 1.06821044,  0.77370361],\n",
       "        [ 0.90852604,  0.96703727],\n",
       "        [-0.54706224,  0.46739791],\n",
       "        [ 1.0576549 ,  0.97139534],\n",
       "        [-0.87353582,  0.87122579],\n",
       "        [-1.23017555,  1.06523311],\n",
       "        [ 0.99854519,  0.68188934],\n",
       "        [-1.42203164,  1.01177337],\n",
       "        [-0.88286982,  0.81583612],\n",
       "        [ 1.23187251,  0.62567337],\n",
       "        [-0.91766462,  1.01195228],\n",
       "        [-0.8894876 ,  1.34728643],\n",
       "        [-1.6845077 ,  1.32857949],\n",
       "        [ 1.32551412,  1.838294  ],\n",
       "        [ 0.88239331,  0.70017571],\n",
       "        [ 1.03345052,  1.46893146],\n",
       "        [-0.93355729,  0.99785516],\n",
       "        [-1.46662161,  0.87874071],\n",
       "        [ 1.033341  ,  1.18559003],\n",
       "        [ 1.08480038,  1.64661853],\n",
       "        [-1.29936179,  0.85335555],\n",
       "        [ 1.17718111,  1.1437713 ],\n",
       "        [ 0.78655651,  0.77528533],\n",
       "        [-1.85064099,  0.89562312],\n",
       "        [ 0.81405135,  0.88883881],\n",
       "        [ 0.93640333,  0.6236295 ],\n",
       "        [ 1.21828168,  0.96457205],\n",
       "        [-0.5789623 ,  1.32720112],\n",
       "        [ 1.15080755,  1.02242019],\n",
       "        [ 0.82550695,  0.55853873],\n",
       "        [ 0.96548233,  1.18468624],\n",
       "        [-1.13742855,  0.62153099],\n",
       "        [-0.84983074,  0.72297801],\n",
       "        [-1.44189096,  1.09488485],\n",
       "        [ 0.70455707,  1.39198128],\n",
       "        [-0.67204982,  0.85877496],\n",
       "        [-1.88873982,  0.44118981],\n",
       "        [-1.15649503,  0.85982375],\n",
       "        [ 1.10018476,  0.98042741],\n",
       "        [ 0.60660431,  0.56998385],\n",
       "        [ 1.00277655,  1.24549398],\n",
       "        [-1.44620987,  0.6675918 ],\n",
       "        [-1.64162479,  0.76170668],\n",
       "        [-1.14722826,  0.88861146],\n",
       "        [ 0.61771419,  1.02817479],\n",
       "        [-0.62373281,  1.25704434],\n",
       "        [-0.94495995,  0.6109426 ],\n",
       "        [-1.53556118,  0.10593645],\n",
       "        [ 1.17353145,  1.14809657],\n",
       "        [-1.21258781,  0.9292506 ],\n",
       "        [ 1.06796368,  0.5630779 ],\n",
       "        [ 1.17899425,  1.17867274],\n",
       "        [-1.73963799,  1.22867367],\n",
       "        [ 0.75148062,  1.06702918],\n",
       "        [-1.36419132,  1.21619061],\n",
       "        [-1.18913787,  0.62368852],\n",
       "        [ 0.85063796,  0.44689516],\n",
       "        [-0.54245992,  1.34740825]]),\n",
       " array([1, 0, 0, 1, 1, 1, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 0,\n",
       "        1, 0, 0, 1, 0, 0, 0, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0,\n",
       "        1, 0, 1, 1, 0, 1, 1, 0, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 1, 0, 0, 1,\n",
       "        0, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 1, 1, 1, 0, 1,\n",
       "        1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1]))"
      ]
     },
     "execution_count": 191,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "make_classification(n_samples=100, # 生成100个样本\n",
    "                         n_features=2, # 生成2个特征\n",
    "                         n_redundant=0,  # 设置0个冗余特征\n",
    "                         n_informative=2,  # 设置含主要信息的特征为2\n",
    "                         random_state=1, # 设置随机种子,\n",
    "                         n_clusters_per_class=1  # 每个簇包含的标签类别有1个\n",
    "                         )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 195,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T08:01:55.090595Z",
     "start_time": "2019-12-05T08:01:54.787540Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x279c5cd4a20>"
      ]
     },
     "execution_count": 195,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(X[:,0],X[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T08:04:58.732966Z",
     "start_time": "2019-12-05T08:04:58.418505Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x279c63a4080>"
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#从图上可以看出，生成的二分型数据的两个簇离彼此很远，这样不利于我们测试分类器的效果，因此我们使用np生成\n",
    "# 随机数组，通过让已经生成的二分型数据点加减0~1之间的随机数，使数据分布变得更散更稀疏\n",
    "#注意，这个过程只能够运行一次，因为多次运行之后X会变得非常稀疏，两个簇的数据会混合在一起，分类器的效应会\n",
    "# 继续下降\n",
    "rng = np.random.RandomState(2)  # 设置随机种子\n",
    "X += 2 * rng.uniform(size=X.shape)\n",
    "\n",
    "linearly_separable = (X, y)  #生成了新的X，依然可以画散点图来观察一下特征的分布\n",
    "plt.scatter(X[:,0], X[:,1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T08:06:40.331622Z",
     "start_time": "2019-12-05T08:06:40.320707Z"
    }
   },
   "outputs": [],
   "source": [
    "#用make_moons创建月亮型数据，make_circles创建环形数据，并将三组数据打包起来放在列表datasets中\n",
    "datasets = [make_moons(noise=0.3, random_state=0),\n",
    "            make_circles(noise=0.2, factor=0.5, random_state=1),\n",
    "            linearly_separable]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T08:06:44.503135Z",
     "start_time": "2019-12-05T08:06:44.473376Z"
    }
   },
   "source": [
    "## 画出三种决策树的分类效应图"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T14:13:54.231202Z",
     "start_time": "2019-12-05T14:13:53.281741Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAKACAYAAADJpewGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXd4HMX5xz+z10/t1LtsFTe5d2NscAEbO8YOPfReElpC/yV0Qg8EQk0goQVCDIRqbIMBG9x7xU2SbXWrtztd253fHycLWW6SrGrv53n02DfanXn3tLvfmXfeeUdIKdHR0dHR0eluKF1tgI6Ojo6OzuHQBUpHR0dHp1uiC5SOjo6OTrdEFygdHR0dnW6JLlA6Ojo6Ot0SXaB0dHR0dLolukAdB0KIt4UQN3W1Hc0RQuwVQvTuajt0Tm6EEA8LIeqEEFVCiB+EEIOOs74XhRAXddTxray3WAjhFUJUNPw/o73b0dEFqtsihHAIIX7fU+vX0WngZSAa+B74WghhaWtFUsrbpZT/7ajjW1lvHLAcuERKGSelzGrvdnR0gerOOICOFJCOrl9HBwAppU9K+RjgByZ3tT06PQddoNoRIcRVQoj3G37KhRCfiAAPCyG+FEJsF0LkCiHObnL8203OXyyEmCSE+ABYAyQ3uA8WHKNdRQjxmhCiSAjxPmBq8rtHhBAFDe1e3lB2xPoPd7yOTjuxCegPIIS4WgiR03DPXn/gACHE3Q33Xp4Q4rKmJze41K9qVvaIEKKwoZ7fteD4A/XvFELMaCh7WAjxVyHEgobn9sW2XmCDe31Kg0vznSblR7rew5brBDB2tQEnIOcBFwA3AlnAsIbygcBIIAVYJITodaQKpJSXNMwhLZZS9m5Bm+cDI4BU4FfAJQBCiBRgItAXCAPWA+8dqf4jHd+C9nV0WkIdECyEGAjcAYwi0JlaL4T4AhgCXN7wrwPYIoT4TEpZd7jKhBARwP8BsQQ62/8AXj1S40KIM4ArgaFAPPC9EGJow6+vAc4ACoEcIcQjUsqKNl7nX4DbCAgyR7neqMOVSyn3t7HdEw5doNqftVLKLwGEEDsJvOgBPpNSVgKVQoj9BESgOaKNbY4HPpZSuoFPhBBVAFLK3IZ5pjsJuFZij1ZJa4/X0WklQQREagqQBvzcUG4D+gFnAe9LKauAKiDkGPVVAzuBvwILCIjP0ZgB/LvJc7iKQIcM4Esp5RoAIUQxEAq0VaCekVIubfL5SNc79AjlukA1oLv42p/sJv9vmom3qfgogHaYcxPb2KZo1pYGIISYCHwK5ABXHbOSVh6vo9NKBgPbCNyv7zYEF8QBScDK5gcLIc4XQiQcqTIppQqMBj4GTgc2CCHMx7BBNvv/gc9Hem7bQvNrOdL1tuh7OJnRBar9OdLN/WshRLgQYjgQCewGaoBkACHETCC9yfHlQKQQwt7wYztKm6uBc4UQFiHEHCCioXxsw+/+A8xsds7h6j/a8To6bUIIYRRC3EfghbyYQETfDCFEnBAihIArLBNYCFwihAhrEKaXOXxH7kC9fYHvGn7uBeIIPFtHYj5waUMEa38C9/uBkU5HbutwpOs9UrlOA7qLr/PYQOBhCAGukVLWNwQn3CGEWAxs5ZeHBSllrRDiaQI9OwU4hcDI5nB8SMCNkEugB1bcUP4xAbdHITAXqBNC9JVS7jpC/Uc8vp2+A52Tj1sIzMeuBKZLKX3AViHEY8AKAu+gF6SUGwEaOnBbCET8/UFKWXz4akFKuUsI8ROwp6HoZSll0VGOXySEeA/YDLgJPIf7hWirZ71lSCmPdr2HLdcJIPT9oDoeIcTDAFLKh7vWEh0dHZ2eg+7i09HR0dHplugjKB0dHR2dbok+gtLR0dHR6ZboAqWjo6Oj0y1pVRRfVFSU7N27dweZoqPT/Vm3bl2ZlDL6eOoINVtkjNXeXibpnKRYzGCJCIMgBy4/lDu9KIqgptqF2di9xx41xTkteo5aJVC9e/dm7dq1bbdKR6eHI4TYd7x1xFjtPDdqSnuYo3MSk9oL0i+ahTJ2FusqBe+sziXYbOSbbzaSHN69O0ALn7qgRc9R95ZZHR0dHZ2TFl2gdHR0dHS6JbpA6ejo6Oh0S3SB0tHR0dHplugCpaOjo6PTLdEFSkdHR0enW6ILlI6Ojo5Ot0QXKB0dHR2dbokuUDo6Ojo63RJdoHR0dHR0uiW6QOno6OjodEt0gTrBKS0tZc+ePWia1tWm6Oj0WDQpKa53UuP1dLUpJxWtShar03MoKSnhut/8hh+XLyfYYMDucPDSv/7F9OnTu9o0HZ0exeqyQt7esQGp+qmTkqHh0dyYOZoQk7mrTTvh0UdQJyBSSs6dPp3+S5dS5PGQ53LxSmEhl597Ljt27Ohq83R0egw5tVX8fdsaPvB5KNBUiqTG0MoSXti8vKtNOynQBeoEZMOGDRTt3s1TPh82QABnAjd6vfzjpZe62DodnZ7Dt3lZ3KmpnE7gOQoCXpCSgrpq8p21XWzdiY8uUCcgBQUF9DcYDvnjZvr95Gdnd4lNOjo9kUq3k4HNyoxAuhCUe+q7wqSTCl2gTkBGjBjBCq+XqmblX9psjJ06tUts0tHpiaSFR/OpOPg1WQJskRqpwWFdY9RJhC5QJyCJiYlcdfXVnGm38zWwFrjVZGJNeDjXXn99V5uno9NjOCsxnYVGE78XgvXAl8BUxcCMxHRCzZauNu+ER4/iO0F57pVXeGvkSJ7829+orq5m2uzZLLv/fhwOR1ebpqPTYwg1W3h89BQ+3bOdCyv2E2wyMzW5D5Nik7vatJMCXaBOUIQQXHPttVxz7bVdbYqOTo8mwmLj2v4jutqMkxLdxaejo6Oj0y3RBUpHR0dHp1uiC5SOjo6OTrdEn4PSAaCqqood69ZRsXcvluBgeo8YQVpGBkKIrjZNR6fHIKVkc2UJWSX5uLxeokLCGBHXi1hbUFeb1iPRBUqHuro6Vn36KQOkZGREBC6Ph60LF+JxucgcOrSrzdPR6TGsKSnAXbSXG0wmoswmttVV8UlWNWf0HUakxdbV5vU4dBefDtnbt5Pm85ESHY3JaCQsKIhRCQnkrlmDz+fravN0dHoEXlUlqySPKy1WkowmrEJhpNnKWZrGltLCrjavR6ILVDfC5XKxceNGSkpKOrXd2uJiIoODDyqzmEzYVBWXy9WptujotAdlbhd766rxdeI2M7V+L+GaRqhy8Gs1zWikxlXXaXacSOguvm7C808/zeOPPkqC0Ui+x8PZs2bx+rvvYrfbO7zt4JgYKouLiQgJaSzz+f3UGwzYbLpbQqfnUO318MrWVeyurSBKKFQCl/cdxulxKR3edrDRRKUQ1GkawU1Eap+qEhza8c/xiYg+guoGzJ07lzcee4w1LhdbamrI9XjwzZvHbZ2UlihtwACyhKCwvBwpJU63m3UFBSSOGIHZrO95o9NzeGHzck6tKadI09il+vlG9fP+zg3srK7o8LYtBiOpMUn82+OmVFXRpGSbz8PXwKDoxA5v/0REF6huwKtPPcWTTidpDZ9DgFfcbj765BNqamo6vP3Q0FBGnXMOOTExzCssZKnbTfjkyQwcNqzD29bRaS8KXLUUOmt4RkoOZMkbDtynqXybt7tTbBgbm4wpMY3npOROj5tPrcGMTx9MjFUfQbUF3cXXDSgqLqZPs7IIINhgoLKyktDQ0A63ITIykgkzZyKl1EPLdXoklR43vYQ45KXWB6jxdM5cqiIEo6ITGBkVj2z4rNN29BFUO1FfX8+uHTvYsn49hYWFaK2YnJ0waRKfNJtYXQkY7XaSkpLa2dKjo4uTTlcipSTXWcOyolxWluS3as+l1BAHO6RGXrPyj4RCRkRs+xp6DIQQuji1A/oIqh0oLS1l/Zdfkuj1EmQwsMfrZU+fPoybOhWDwXDM8//v0UeZ+PXXeJxOZvn9bAYetFi47Yor+PF//yM6I4M+mZlYLHp6f50TFykliwty8JQVMU4I3MAPxbkMSO7DwPDoY54fZDRxTq/+TN63k0c1lWTgDQTfKAqT6+v4MmszfaIS6BcWqXfEegi6QB0nUko2ffcdIy0WoqIDD1GqlKzZtYt96emkpacfs46MjAyWb9jA8088we9/+onw8HCeHjOGWSNGYNY08latYllODhNnz8ZkMnX0JenodAm5zhq85UXcYbVhahCQ0arKM3lZpIWEYzMe+3V1Tu/+JASF8re83VR53TiEwv1mK6OBareLr/buoCa+N2NiO9czodM2dIE6TmprazFUVRGV+EuUjhCC3mFhZO3e3SKBAkhNTeWlN97A5XLx03vvMTU2FmPD6Gug3Y4nL4+83NwW19cZqKrKv//9b+a++Saq3885V17JNddeq4uoTpvIrankVCEaxQkgwmCgn89LvquGPqERLapnbHQCY6MTWF1SQHTRXmZZA0slYg1wg8HIn/fnMSQqDquh+7z+yj31fJ27m5yqUiJtQUxL6UvfFl7viUz3+Qv1UBRFQYVDggv8qoqhDS/q6upqIqBRnA4QY7NRUlwM3Uigrrn4YnZ+/TV3Op0YgZc2b2bexx/z2TffoCj69KZO6zAoCm55aLkHcIjW30+VddWc1uw5ClYUkpCUe+pJtIcc4czOZX+9k/vXfs9vVD83ScmWumqeKN/PdQNGMS7m5A5P198ix0lwcDDmpCRyy8oay/yqSlZdHYn9+rW6PpvNRq2mIeXBT2qtx4OtE6L5Wsq6detYPG8ePzidXACcAyx0udizahXfffddV5un0wNJD4vkJ6BaUxvLsv0+9hpNpAS1/t63W20Uq+pBZX4pKZGSYGP3Wd/36Z6fud7v429ScgbwB+ATTeW9XRvR5GEU+yRCF6h2YMSkSWSFhrIsP58NBQV8X1xM+PjxbYrAczgcWNLS2F5UhNrwcO2vrCTXbCYlLe0YZ3ceS5YsYY7fT9M8EybgvLo6lvzwQ1eZpdODibUFkZaUzlM+H/921/MPdz3/kILTUgdgbMOIvH9ELN8IQbbPi5QSt9T43O0m2BFJmLn7BBz9XFnKxc3KJgJe1d+qKMYTEd3F1wL8fj9CiCNG5AUHBzP53HMpLS3F4/HQNzKSoKC2p9cfPXkym1et4puff0bRNCzx8YycMOG46mxvoqOjWWw2g9d7UPleq5XhsZ0b0qvTM5BS4tVUTIrhiCHYQyPjSA+NIN9VS6SiMCIoFJNy7EjYwxFttTMmbSBv52cjPPV4hCA+MpZJCanHcxntjsNkZp/XzaAmZdVAvZQEGU/u+VxdoI5CbW0tW1etojIrCwwGYjMzGTRq1GHDvRVFIbadXsxms5lREyfiGzeO3NxcSrKy2Pbjj0SlpZHevz9Wq7Vd2jkezjnnHO659VY+As4HBLAA+Mpg4NlLLula43S6FVJKtlWWsK04F83nRTGZGRCbwuCImMOGewebzPQPi2yXtnsHh9Gr33DK3C52VJVRUVvJor3bSY+Kp19oRLcIN5+c0pd7dm5guKaSALiB24XCuKg47LpA6RwOr9fLii++oJ/Hw9j4eDQp2bVlC6uqqpg4c2an3Ng5u3ZRvHgx/UJCsJpMFKxcybKsLCbOnt3lOfKCg4P56rvvuGTOHP5UXY1RCOptNj6ZO5fo6GOvWdE5edhRXU5+7m5uNZuJt9ooVv28l7cbRVEY1IL1TceLV9NYkruL4W4nc4wm6rxuFu7ZTlVcCuM6IYnssZgUm0yZq5YBebtJFwr7pEamI4qb+4/satO6HF2gjkB+Xh7RNTX0aphHUoDMxESW7NtHeXk5UVFRHdq+1+tlz4oVTI6Px9IQDegIDsafn8++PXvo04YAjPZm5MiRbM/NZfPmzaiqyrBhw1q0MFnn5GL7/lyuNpmIbwjrjjMYudgkeb04t1MEakdVGf3cTs5tkg+vt8HIo/vzGRQZR7Cpazt7QgguSBvIjOQ+AdemxUa0nrsP0AWKnJwcvF4v/fr1O2hU5KqpIewwCwMditIpeyTV1NQQ6vc3itMBYoOC2JeXB91AoCDg2hymJ5U96fGqKoX1dTjMFhzmg13QdR43ic2CEhIMRpwed6fkfix3VjO52TyWTVFIF1DqcXW5QB2gPV2bJwonrUDt2LGDK887j9w9e7AoCtbwcP754YeceuqpAIRGRlLs89E0bk5KSZmm0asTwr2tVitODl1fVed2Yw0L6/D2dXRayvy8LObu2UYUgv2axsjIWG4YMLox84PDFkyWx8WAJkKQ5ffhsAV1iqvcarZS0iw3ppSS/VLSqxuFm+scykkZZu7xeDjr9NO5evt28uvr2eN08kx+Pr+ePp3i4mIAEhMTcSYksC0/H6fbTY3Lxbq8PIIzM3E4HB1uY3BwMCF9+rCtoKAx3LyitpYcoHffvh3evo5OS1hdVsSCnG2sUFV2qX4KpEZ8+X7e3L628Zgh8b34UFXZ7PVQq2ls9Xr4QFUZEterU2zsHx7NYqE0hpv7pGSBx40WHEa0Rd+QsztzUo6g5s2bR1p9PTc1WQQ3G5ijqvz7vfe46+67MRgMjJ8xg13btrF8+3YMJhOJkyeTcQTXmqZp7Z49YcTEiWwymfh2+3aMUoLDwZDZswnTR1A63YRv9+3kKU3lwFMRDLwmNZLKi6nxeQg1WegdHIYhYzBfFedR63YSYg9hRGwyvYIPvY+llGiAoR1HVhEWG+PSMnkrPwuTx40bcIRFckZSereI4tM5MielQBUXF9PH5zukvI/bTVHeL8n6zWYzZdXVrM/PJykpiXG9ex8SBFBaWsr2FSuoKSjAHBJC75Ej6dO//2FvfI/HgxACKSVPPPII7775JnX19cyYPp3HnnuOXr0O7lGazWZGn3Ya3nHj8Pl82O12/YHS6VZUet00H8+HAuGKoNbnJdQUmHuKtNiw2IPwGI30cUQfkhlC1TTWlhaQXVqIT/URERTGyIRUEuzBh7Spahr1qh+rwUiBq5a5WVvYWl2Ow2hianIGs5L7HLLOqldwGCn9RlDt82BWDCd9+HZP4aQUqPHjx/OUolAPjZkQNOCz4GDuOv10ANxuN+dMm0buhg3MdLv50Wrlvttv59ulS+nfvz8AlZWVbPj0U4ZarcQkJuJ0u9n83XeoPh8DhgxpbK+mpobNy5ZRt28fEnjpX/+CLVv43O0mAnjj00857ccf2bhzJ+Hh4YfYazabjxhWvnDhQp598EGycnIYPGgQ//f444wfP779viwdnaPQ1xHFx8W5jGhStglwCYVYa2Bh+c7qCp7etJQJUpIqNV4WCukRsdwyaGzjSGlZ0V7spQXca7biMNrYWl/Hh9lbOLPvMCIa3HBSSjZX7Ofn4lwsqp9qKfmxspQHpMbHwD6vyu17tlPpdnFl30MDd4QQhwRwHMCjqny+byfLi/ehSo3RMUmc23tAtwmgOFk5Keeghg0bxulnncV0u515wPfAeTYbpr59mTNnDgAvPv88xrVr2VRXx7N+P/Pq6vhjRQXXXnRRYz3ZW7fSz2AgxuGgqqqKssJCYvx+di5dit/vB8Dn87Hyyy9JKSpiekIC/X0+zBs3crs7sHI8CXhE05hYV8dbb77Zquv45OOPufbcc7lu9Wq+Lyvj14sX8+szzuDHH39sny9KR+cYzOk9gDeMJu5FsBJ4C5ipGLg4fTBGRUGTkr9tXcE/VT+fayovSMlOTaWuYj9LinMBcPl9FJYVc5nVRqgQVHjcRPu8jPO42VRa2NjWzupy9udlcaeARyxWLq2v4zyp0R+IA8YC8zSV74r2UuP1tPgapJT8ZdNSavJ2MddTz5deD8EFe3h0/WL8rdh4VKf9OSkFCuDtuXO59LnneHb4cO7PzGTcAw+w4KefMDZEHs196y3uqa8/aIh5vZTs3LWLgoICAJylpYTZ7ezeto3CFSuwZmVh2bWLvOXLycnOBqCgoIComhpSoqMRQpCXm8spRiMhQEmTuqfW17Np5coW2y+l5IE77uBdl4vfAGnAtcDz9fU8evfdx/PV6Oi0mFhbEE+Mnsr2+F5cbQviTUc0Nwwex9SE3gDk1FVhU1XmNDnHCtypqawu2gtAjc9LjABN1dhbVY7JWUO020WC28nGor24/AF3/I6SfM4zmYhuWE+l+HzcDHwHHJhNjgQyhEKBq67F17C9upyK2io+1TRGA0OBt6VGlKeelWWFxzpdpwM5KV18AAaDgRtvuokbb7rpsL+XUh6i3qLh50Cm8ZC4OLKWLsW+dy+ZDQJU7/ORKCW7ly0jPSODeqeTkCbBE/Hx8WRpGjOApqup1lgs9Bk8uMX2u91usgsLmdysfAZwy5YtLa5HR+d4ibHaubb/iMP+7sBz1Hzm1MAvohJmslAC7HPW0BdJdMM8ryohTZOsLy1gQnxvnB43cU3WJtpMJjS/l3rATyBZcR2QIzVibS1f6JpVW8U0KQ96GQpglqqyprqCCTH65oZdRY8dQbndbvbt24fH0/KhfGs474oreN5qpekA/x0gLS2tMUt5+sCBrKuowOfz4dc0KuvrWVNZyaBBgwj3eqmoqCAsPJySJin/MzIyiE5O5sWGfaS8wJvAJ2Yz1954Y4vts1gshAcFsbNZ+Sagd3x82y5a56SkxuuhooOyZqeFhFOrKMxvUuYFnlcMjGoIM7cZjSRExfO+x40LcErJUlVlsSK4KCiYosrAVjYRwaHs9P+SnDg5OIwFBERJAPnAJYrCqMj4xnmrlhBjtbPpMBG46xUDMYcJ0tDpPHqcQGmaxsN//COJkZFMyMwkKSqKpx577JD9k46XP9x1FxWDBzMqOJgHgfODgrjf4eCN//yn8ZiwsDD6TZ5MjsPBoro6NppMpIwaRd+UFPxSYjAYiIuLQ+vdmw25uVQ7nVTW1THt+uupnDSJ4WYzIQYDH4waxcIffyS+FcKiKAq/v+surrHb2dNQtg24xW7nD/ff367fhc6JSYnbxWPrFnPz8vncuXIh9636hqyaynZtwyAEtwwax6UGAxcrCn8EBhoMGBxRTGmSB29sbDJ5NjuvCcH9UmOt2cLlYREECQVDg3gMi03hYwmrPG5KVZUsKVkXEs6PZit2BJmKgiGuFzcMGNUqG0dFxpFnMPI4AjfgA/4OLFYUTotNbrfvQqf19DgX33NPPcXCF19kvctFLyALuOCppwgLD+e3t9zSbu3Y7XYWrVjBwoULWbVyJdOTk/nXRRcR2iyLxMARI9iSnc2EuDjMDWmJiioq8EVFER4ejhCCU6ZNI2vnTtZv345iMJA4Zw4f33UXqqqiqmqbs5Pf86c/4fV4GP3CCxg0DcVs5o8PPcQVV111vJevc4Kjahp/Xr+EGzxu7kJilPChq47bNv7EX8dNb9f9kgY6onjxlLNYuj+fHK+HK8OjGeSIOmjJhMVgZHBMEpmVpZxlsSKEwC8l77pd9GpwscXbgzm9z1B+KslnnquWoOAwpqQN4oqgEDyqH6NiaNP6KaOi8MDISbzx8xqerKlAIEgLCuGBAaNO+u0uuhrRmpHHqFGj5Nq1a499YAchpSQpMpKFlZUH7Z2yArg6MZEd+fldYtfObdvYs3QpMVLiAeocDsbMmNFpC2q9Xi/l5eVER0c3BnnodAxCiHVSytZ10ZuRERounxs1pb1MahNry4pY+PMaVqv+g8qvUBSMqZnMSen8bCUuv49v9+7A7qwhEdgFBIVHMyUpo3EU1dHU+bxoyMb1W92Z1F6QftEslLGzWFcpeGd1LsFmI998s5Hk8O6dbHbhUxe06DnqUW8zTdMoqqwks1n5YCC3pORwp3QK/QYOJCUtjbKyMkwmEzExMW3KKlFTU8OuTZuoysvDGhZG6rBhJCYmHvM8s9ncKvegjk6pp57B8tAQ6qGaxk/1zi6wCOxGE7PTB1FYX0etz8upbczqLaXk56oysssKcfu8RIdGMDwm8YhroJqir3vqXvQogTIYDAzv04f5u3fzqyblXwFjWhEB1xHYbDaSk9vur66trWXFJ5/Q1+8nMzycmrIyvvzLX/CmpXH2nDmkpHT9vjU6Jw59QsN5AYEXOPBKlsDnBgOjHB27lczREEKQaA85rjpWl+TjKtrHJSYTDqGwvqyIf5fkMywxnbHR8W3eoVen8+lxQRKPvfAC19psvAXsAP4B3Gaz8fBzz3WxZcdH1tatZPj9pMbF4ayp4Yn77mPNa6/x9d13M6xvX+68+eZ2DwTROXnJCAmnlyOSXykGfgLWAlcIhf1mG+OjE7ravDbjVv3k7M/nOouVdKMJt7seY2UJw6rLmbd9LTct+5qfq8q62kydFtLjBGrmzJnMXbCA/51+OrPj4ph/xhl88f33TJo0qatNOy6qCgqIaQjAeP6JJ5hYUsJLHg83ejxs9nj44Z13eP/997vYSp0Tid8PHk9S7/5cZw3iQosNT1I6D46c1KNHGJUeN/FIghWFOp+P3dVl/Bq4BLhKanzg9/GXzcvxNJt70+me9CgX3wFOO+00Tlu8uMva1zSNgoICyvLyMNlspKSnHxLd11rsERFU79tHfV0dubm5RGsalwAbAQdwntPJOy+9xGWXXdYel6Cjg0lR+HWvfvy6V9dtflnj9bCjqhSnp56IoFD6h0VhOY5dmYNNZkoattQoctWSRCDTxFLgbeA8YJCUrCvfz/iYY8/v6nQtPVKguhJVVVn53Xcou3eTaLPh9vtZuWYNmTNmkHQcc1BpgwezcedOYmprwe9HADcD0QSimW4Aksta7prYv38/RUVF9OnTB6PRiMlkavftQHR0jodCVx0/Zm9loqYSrwi2Vezny9JCZqYPanO28RCTmciIGD4q30+634cTSCSQueJH4AUgT1MZqh66m8HhkFKS76pFEYIosw0NGjdi1Ol49G+6leTl5mLcvZsxycmN6zjiXC5WfP898Zdddsh2HC0lOjqazLPPZsEHH7BJSi4hkACzP4EoxVXAyshjbwftdDq58fLLmff118QajXi8Xs7/1a8445xzSBszhj4Nmdh1dLoSKSWr87O4XEgyrYGsD8OBL+pdbCwtZHx82zczPC0xjZUGIx/lbCOeQBLZa4CBwOtADBDXkGn9aOyoLufVbatRvR40KYlUFNKDHUSGhjM2KZ1Y27Hr0Dk+dIE6CgdceaW5uZhsNpLT0ijZu5eU4OACuy5IAAAgAElEQVSDFhmG2u0EVVZSVVVFZAtE5EgkJiUR3b8/cTYbM+rrD5ogHA3sCj522pVbr70WOX8+yz0e9ns8JALPz5+PITWV4poaDEYjaRkZbbZRR6ct1Hg9bK8qwemuJzwolOSgULz1TgZYDg79HmM28WpVGRyHQJkUAxMTUvmpaB9X+Cq4iF9yAZqBTKHgO0yIfXN7n960jH+qfrYDfYG+msr6mgrCTCbmZm9lVr/helh6B6ML1BHQNI2V33+P2LmTJJsNj9/PmtWr8UZFEes/dILVK2W7LJIdNmwYG6TECTQNtl1gtTL6GIEg1dXVfPLZZ+zzeNgIDAEigCs8Hr764gvuffxxVq1ZowuUTqdSVF/HkuytTFBVEhWFHRUlfGu24JEaXuCgJbE2SaTNSGo77Aafmh3FD7VV/KaJGFUBPwu4Pujoc8ZL9ucyU2oMIZDf8jwCIpcrIFJVGSMlO6rKGNWDIx57ArpAHYH8/HyUnTsZ28SVF+92M6+ggF2KQpzPh6UhtdG+0lKUhIR2yRyRkpLCBRdeyKyPP+YJl4tY4B1F4dvgYNb99rdHPbeiooJQgwEHgUzpBx7BeKCispIQmw1PQQFSSn1nXp1OY3V+DhdLyZAGV94wwOGu53ODiUUeNzMbUhuZ7ZL5Lh9n3nAx6cObL8dvHX63hxtcbi4u3EO8X+MKoAC4RzFwWmzyMZPJVnncDNM0ygjs2XbgaQmTEo+mkmgwkO9xH5eNOsdGFygCGRzycnLwOJ1EJiWRlJREyd69JAcFHfQiD7JaSbZYEIMH88PPPxOhqrilxB8by5ipU9vNnlf+9S9eHDSIm155heraWs6cPp2fnnyS6Ojoo56XkpKCsNlY7XIRRmC/qQRguRBkDh5MWU0NIQkJujjpdAiqlGTXVlJQXYHRYKRPeDThZgtOVy2DmrnyRpnMLNY0NgSHsamumnS7IKteMu78GZx+w13HHdBjXPUVIy8+m7dUlRd+3MrrlSWEGE1MSspgZlL6Mc/v74jio8IcrlJVPiaQQFYAuQgGmS0s0TTC9UznHc5JKVBSSlavXs3OnTuJjYlB5ObSGwg1GincuJHc1FRsDgeew7jyPFIyZMAAgkaNoqKiArPZTERERLu+9A0GA3fcfTd3tHLjQYPBwNMvvsi5N9zAHS4XmwCPEKyxWrlrxgw21tUx+Iwz2s3OnkB1dTU5OXm4XF6SkiJJSkpqcyCLzqG4VT8bKvbjU1UqnbVE1FYyQVFwSckPpQWkJKQhhcAjJbYmz4hLalhNZmalDcQS6cJ+ynB+kxRPzLSLARD2ti/bkK4alLGzMK76ilMv+zVJIUHs2de6OkZExvF1UBjX1VUxSNN4CEgBrGYLP6oau2125oS1fb65pyE1leqibKoKizCajET0SsceHtfh7Z50AlVdXc2vzzyT/J9/ZgywwetlTno6Dz/2GBaLhWRgzZ49MGYMOT4f8R4PNkvAS55XWoqMi2vMUt4d899dfOmlJCQl8dKTT7I3O5v+/fox68wz8Q4cyLChQ485CjuRyM8vYN68bQiRisnkYOPGAtLScpk2bbwuUu3Ahor9vLh1JSNoECBN5erQCJJtAXEZoqk8VbSXqLAoFlSVMMdiQ2kQq699PnrH9SKtt6DXnF9jtFpQxs5ibbmGUAxQWXsclgWEcOTYWZj8XtKtFvjvV60SKYMQ/N+wicwvyObbon3UqyrJVhvx1iASwqP5VXQC5pPkHpKaSs6KRVTmh6EYTwPNyf5dP9BrdCpRqYOOXcFxcEIJ1GeffcZrTz9NcXExE6ZM4Z4HH6RXr4NnW+++5RYyNm3iO6+XagILYVdnZ/PBO+9w9Q03AJASEsKeykrSzjyTJYsXE+7349Y01Lg4Rp9xRrd3kZ1++umcfvrpXW1Gl6KqKt9/vxWH4xTsjb3xJHJy1rB37z7S09O61L7ujNPv44t9O9lYWoBZMTA+IZVpiWkHbWXh9Pt4YctKvtJUJhDYzDMKyK6pJMJqI8howqEY6CO92MMi2KH52VJTSZKAbAmx0Qn8ZngsveaceZA4vbum4NDtd9uCBMamMDLcDCOnkQ7w368AWixUZoOBOSl9uySze3eiujiHyvxQrKE3IUTA9ar5h5C37ikciRkYW5CEt62cMAL14nPP8fKDD/K4y0U68PG77zL+009ZuWlTYxJXTdP4YO5csr1eFAIXrwKX+nz86YcfGgXK6/djtFpJy8ggKSWFiooKTCYToaGh5OTsZffuTZjNBgYMCMxXdXfBOhmprq7G5bITHn6wqygkJJmcnL26QB0Br6ry8NofGOV28S+pUQs8nL2VPdXl/G7gmMbjVpcVMVEIJjR8thLQlX5Iilx1ZISGA4HdbmNMZmamZlLqdlHr83KWxYZHU/nv9mKqq5bSd8JYZvWpQShhICDYfPyvpTpvwD0v7KHgqkECjr7pVO3KPu66TzaqC4tQDBMaxQlAMTrQtHRcFUWExqV2WNsnRGoBl8vFow8+yAKXiwuBkcCTfj8X19XxwjPPNB4npcTj9zdGt4UQCHEtA9zewFbSHp+PLLebpL6BXpPZbCauwa23aNFKFi2qobq6P8XFqXz++V42bNjaiVfaPVizZg0XzJhB/8REzp40iSVLlrR7GxUVFaxdu5by8vI2nW80GpHSe0iCXb/fi82mb0J3JH4qySfJU8/7UmM8MB34TlNZX1ZIvvMXt5tX9eNo8t2OAxYDTkBrKN/i9bDfbGnMTh5ttZMW4qDO7+ONXU6WF55NadVVLPy2N3fc8TKFuXs76Sq7B6qUfJ2fzf+t+oY7ls/n3V2bqPF62rUNKSXOikJq9u9Bamqb6jCYjUit7jC/caJ08DqwE0Kgdu7cSaLBQPPYnDk+Hyu+/77xs8Fg4MxTT+WNJiOekQRWl1v692d1QQE/lJeTOHkycXEHTwAWFhayZ4+RlJSRhIZGER4eR1LSKaxaVYjT2TX757SGwsJC/vLss9x3990sWLAATTv6QsUjsWzZMn41aRKTFizgf4WFnLtkCb+ZOZOvvvqqXexUVZXf/e4OEhPTmTr1BpKSMrjuulvxHyZg5WiEhoaSnGyipGRvY5nX68btzqJvX30b7yORXVnCeZp6kJfNDkxFsLOmorFseGQcXyEpavjcC5gK3A58aTDwV3c9/zGamZKaecgut98W1SDEhUTbRxJiTyYh/gyQ5/P1Jws79uLaASklWytLeS9rCx/t3cH+49g7643ta9mSvZW/uer4yFNPVOEeHlj7PfWtvNePhLOikGVvPsDyfz3J6vf/zg8v/Z7SnA2tricyJQNYjOr/5e/vdW3GElxKUETHzsOfEC6+uLg4Cnw+XAQepgNsBxKb7aP03N//ztTx49no8TCuvp4ldjs/2e188e67xMXFMTQ8HIvl0N00i4srMJsP/mMYDEaEiKaiooKgoCBqa2t56623WbRoJampCdx88/X07fuL/3r58uXcddcjbNq0hri4FP70p99z9dVXdriLcNGiRfxmzhzOVVVSPB7uff11Xh07lk/mz8dkat1o4uE77+QvLhdXNHzOBOJdLu697TZmzZp13LY+8cSzvPPOetzu3bjdUUAF//nPhcTFPc6f//xQq+qaPHkkixatIS9vH4piw2CoZMqUdGJiYo7bzhMVhy2IbUKBZpkWdghBZpO1QzFWO7NT+jEydxe/01RswN8VA72i4khL6YdZMTDJHozS7N6WUrK3TiXKcnB3MjJyOLt+fpvMUwOfy3J3s2nBZzjLq+g9YhgDJs3C1NC+pqqs+/w9Nn79KV5XNYmZYzjtqpuISOpYt60mJa9tW012eTGXaSplQnDvvp1c138EE2Jb1+kpctWxurSAfZrGgWD1v0uNs31efije16JQ+KPaqqms+eAZPHX3Ab8DFFQWs/F/53LqdY9hd8S2uC6bI4beY9LJXfcMvvo0wIklqJyMCZMOcvt1BCeEQMXHxzN1yhRu+e47XvR4CAHWA4/Z7bx3770HHTtgwAA2797NW//8J6s3bWLM6NG8cvXVOByOo7YRFGTB7z+0tySlE6s1mfLyckaOnEhpaSYu12yMxh28+eYE/ve/95g+fTpr1qzhzDPn4HI9C/ybnJyt3HrrrZSXV3H33b9v1fUWFRXxwjPPsHzRIuKTk/nt3XczefLkwx7r9/u56qKLmOtycWCT8Xvr6pi2YgXvvP02111/favaXrVxI3OblU0Dzt63D4/Hc1hxbw0vvvgaLteXBKbcASJwuV7m5Zcnt1qggoKCmD37dCorK/F6vYSHDztu+050Jsf35p7c3cyQMIvA+p+/IKg2mRkUfnAE6LmpAxgUEcvS4n34NZVLYpIZFhFz1A6XEIIQo8CrVQG/hGnX15cQFh5wBe5a/i3fvPQUqv93SC2N3C0fsmHeZ1z89D8w24L47vVn2bk0H793LpDM3o3vU3jfDVz21/cIjW5dj35TRQmLcndR7XXTNyKGWSl9j7jz7tryIvLLi9miqYGOsJTcIFVO27GekZFx2FqR4DartpJTgeYrqc7RVD6sKoXjFKjyvZvxe+OAW5qUTkLKK8jftIS+p1/Yqvoiew/EkZiBs6IIgzEee8TEDhcnOEFcfAD//PBDPGedRbLFQlpQEGc7HDzx6quH3ScqKiqKu++9lzc/+IDb//CHY4oTQK9eyZjN+VRXlwKBnuD+/TnExfmJioriySf/QlHRBFyuj4Er8PufwOX6gKuvvgVN07j//qdwuR4FriKQo3wyLtcnPPbYk3gb5r9aQkFBAWOHDMH96qs8tnUrZ86fz5WzZvH2v/512OPXrVtHpM/XKE4AJuBml4tP3323xe0ewApsa1a2C3DYbJjNx++Prq7eDzTvCadRU7O/TRs2CiGIiIggLi5OF6cWEG21c9fQU7nVYiNBMRArFD4KCeOPw087xFUH0Dcsgmv6DeeGAaMYHhnbIm/AqTEWqr3z8Wv1AHi9NZSVfcjkmaeg+n0seu1p/N6vkdojwJX4PV9TW9qPTfPn4qwsY8dPX+P3fkkgQ2UcyDvx+y5l/Zf/bdW1fluQwz+2rOCKyhJecNYQnZ/DfasXUXmEDBFr9+dx8wFxamAoMEIINleWtqptgI2aRvM7ej3gaEEi22PhdVaDPFTkpJqBu7ZtIfwGk4XQ2N4ERSZ2ijjBCTKCUlUVj8fDy2+9xYuqSkVFBampqa12Xx0Nu93O2WePYPHizeTnA6ikpgZx2mljEULw6acL8XpfbXbWVGpqvGRnZ7Np02bgiWa/74eqmiguLm7xlu7PPfkk51dV8XyDn3oSMNblYtof/sAll112iEgYjUa8UiI5OHrX2/C71rB79268qsrNwP+AdCCfgORmpKe3i6ty5MiJrFr1cUOtB/iYYcMm6tGSnUCN10OIycyfR02hTvVhUQzHTAvUWkZHRVGv5rCx7hk85QMwCjOXX3E6aadOYtmHC0DGEhCfAwj83uvYtfwR4voMxGAahOo7OK2Y5j+T4t1PtdgGr6ryQfYWlmkqAxrKJksN4ffxVd4uLs8Ycsg5ilA4XAiDBzC0MvNFTlUg+OcB4E8EOn5fAm8Bf3Ac/wJgR1I/pPwAqOGXpGcSg+lDolJHHnf9nUWPF6j8vDy2fvcdIW43XikhLo5RZ5zRruJ0gOjoaM4/fwp1dXUYDAbs9l/6UqGhIQTiAZviwe+vJSQkhIyMDPbvXw003RwuDyldrVo8+9M33/C3ZpOoQwCHlOzatYtBgw5eODd8+HD8oaF8XFfHBQ1ldcDzQUHc2RBW31Jyc3MZarfzq5oaxhGY76sl4OKrDmqfrQdefPHPTJkyC7e7EE2biKIsx2r9C3/726ftUr/O4fFpGksKsqmqKCFRCFYgiY+M59SE9g8hVoRgUlwsv0n0Ez5tAFFTL8AeGcvaglpMtiA0rYLAApCmC2FLsQQFERabhOrbDtQDvwinUFYRmZzUYhvyXbXEIhrF6QAXScnN5fvhMPmUx8f34qXSQq7U1Ebn5GJgB/AHR+sWwFe4ndwLLCCQK9NKwKk9UDGgtd5RcAhB4fHEDxxP8c8TUX33A6EoxlexOSqJ6z/u+BvoJHq0i6+mpoaf58/nVJuNUxMTmZyUREZlJasXLmyTO6glCCEICQk5SJwAbr31Kuz2RwjkSwaQGI2PM2rUGOLi4njkkbuw2+8DFhFYRbgLu/1Sbr75d9hsLe+hxsbHk9OszAWUeL1ERUUdcryiKHz4xRfcFhbGr4KD+Z3FQn+7nZHnnstFF13U8gsHhgwZwmavl6sJjJwWN/wbbrEw5ghzYK1l7NixrFmzhIsv3kNm5n1ceOEuVq78ngkTJhz7ZJ02s6Ykj9jy/TxosXKTxcpDZiuW0gI2lhUd++Q2YjMaSYyOOOj+dyT0JiwmGqE8D40OsFKMlscZNvNXhMYkkDJ0HAbz5QTSv6rAXAymlxk5u+X3c6jJzH6p0dyZlw04jjBiHOyIZkxiKv0UhWuFwhzFwDmKgd8PGtfqrBKpjih+UhS+BLKA1cAyYDeStJBjTzm0hIFnXUnmWZNwJD5LSMw9pJ8axtjL/w/F0HOWWfRogcrbs4deUhLSRCySo6Mxl5ZS1ordZ9uDq6++iiuumIDVmk5o6ByCg/vTt+9CPvroLQCmTp3Ke++9TErK7RgMNkJCJnDnnWfy1FOPtKqdm+6+m4fsdrIaPruBO81mphwmNP4AI0eOJLuwkEv//nf6P/MMX69Ywd/ffbfVCTmjo6O56be/5Sy7nR8IrHl5zGBgXnAwv7vttlbVdTQyMzP597/fYNOmJTz99ENYLBZKS1vv4z9ARUUFDz/8Z8aMmcY551zGjz/+2G62nghIKckpLWS2xYKpwY1qEYJZZgt7OlCgjsTs/3uCkKg3MVkHYbbNxGDqw7AZk0gfE5hJ/dUdDzHgdCsG0wCEYicy+UnOeeA5IpJaPtqLstrpFxrBnUI0itRO4CHFwNTkPoc9RwjBJRlDeGz0GSgZg+nddxivnDqToRGtjwo9IyGVpUYztwtBIYF53GmKgQkxyURb7cc6vUUIIUgYOJGxl9/DiAtuJiQmmaqCnfjch1vT1DKqCnez8bPXWPHOU+xa/F88zqpjn3Qc9GgXn8/tJvgw8yhWIfD5Wralc3shhOC11/7KH/94B2vWrCExMZExY8YcNG9y7rnncM45v6a+vh6r1dqmjM2zZs0i57HHGPvAA6QYjeR5vUyYMIG3//Ofwx4vpeSTTz7hn3/9K5UVFZwxezYJCW3fw+aJ556jz8CBPPL881RUVjJl+nSWPfoosbEtD1ttCdXV1Xz11SqqqyMQIhgpt5GZaeO000a36nsrLy9n2LDxlJWNw+3+PWvW7OGbby7l+ecf4sYbr2tXm3sqmpSomkqQOHj+MlQoePyd+xwBhMUmcfUr/6Vw50bqqyuI73cnQU0iCI0WK2fcdA9Trr8Tze/DaGlbqp2bB43jla0rSaypIE4oFErJRWkDGR55+Hu51ufly9xdbCsrwm40cXpyBnZD216hwQ3zfP/b8zOzy4uxG4yclpTO9MT2D5Uv2rGWws0FSDkGcKMYFpB6ylDCEw8vxEesZ/sKts57D81/HzCA2pJPyd90P+OveQRrSMckzu0SgSoqKuKnn34iIiKCSZMmtXmjv6jERPatWUNKk/2NPD4f5YrCkOPY2fZ4SE5ObkytdDiEEIe4B1vLbXfcwbU33sj27duJjY09anuPPfAA/33hBR50OkkA3svJ4dQPPmDVli0til5sjhCCa669lmuuvfY4ruDYLF68EZ9vIMnJiQBI2Y+tW1eTnNy6NEUvvPAypaWn4vH8EuXock3hrrtO44orLm2Ve7W7IaVkZ00FpW4XaSGOxowNrcWgKEQGO9jsqmW4+ZdIx/VeL7Hhh7qNOwOhKCQOGHHUYxSDAeU4EraGmMzcN/w0St0uqr0ekoNCsBxBcFx+H/ev+Y7JXg+vSo1i4KHt68ivqeQ3GYPb1H64xcq1/Y9+jceLq2o/hZuLMQfdg2IIzBOrvonsXfk8IWcnYTS37P7XNJXtC99H839BIG8ISHUGfo+dnOXzyJx+xdEraCOdLlB/fvBBnnvmGSaZzRQA5UFBfLFoEQMHDmx1XQkJCeT278+qnTtJDgrC5/eT4/WSPmXKCR9SHBQUxKhRo456TFlZGX997jm2u90ccP5N9Hq5tKyMf7z+Ovfcd1/HG9oGnE4nRUXeg0Z6QggcjjR27NjVKoGaN28xHs8fm5UOQFGS2bJlC2PGjDnsed2dKq+bpzcuxet2MhjB21JjRGQ8N2WObnVEGcDohFTmZm9hv9tNikEhS1VZYTQxrZULUHsi0Vb7Md1qiwr3Msrn5a0mC5gnaSp9CrI5K6XPEddOdTXVhXuBUxvFCcBgisFXn0ldaS6OxH5HPLcp9VUlaKqJA+J0AKldRlnOxe1ncDM6VaAWLlzIO88/zw6Ph1hPIGDzrdpaLpg5k21797Y6jFhRFMZNmUJ+374U5uRgtFgYnJHRbbaUyM/P5/33P6C8vJKzzjqTyZMnd2qo9Pr16xlhsRDnPngq+Fy3m3cXLoRuKlBHo7VfX2xsFJDbrNSHz1dIZBeNstuDN7evZYarlr9IiSAQKDOjvJh5+VnMbkP27VhbEDP6DmdbxX62uJ047CHMCo8huINzrbUEKSV5W1azb+MqrCEh9D9tJiFHcMN1FLsrirmlWS67GGCkUMiqqWRUVPfbegdoWFtypICxlj9MJqsdTasmMPPcNGJ3H6Y2jtxbQqcGSbz72mvc6XTS9Na6CpANiUHbgqIopKSkMHrSJIafckq3EacvvviCvn2H8uCDe3j2WSuzZ9/MOedcgqq2LWFjW4iLiyPb56N51r3dikJ8C9ZdlZWVsXndOjatWUNJSUnHGHkYgoKCSEgwU15e0FimaRpVVdn069e6F8Ef/nA9QUGPQ2Psox+j8SEGDx5IevrxrdbvKpx+H+srS3mkQZwgEPL/Z03lp4I9ba43zGxhfFwK03oPYExMUrcQJ01V+eLJe/ny6edY93kyKz908s4tF5Gztv0TFB+NMKudnc1e6BqQIyXhx5gD86oqmytKWJyfxeqSAmp87ZsQ9mg44lOBZWjqL4ERqm8/iuFngqNbtvYSwGwPI6LXEIRyJ4FVlAAFGEx/InXMpPY0+SA6VaDqqquJaFYmgHBFoa6u7ZEl3Q23281ll11Hff18vN7XgIdwOjewaFE2H3/8cafZMWTIEJL79uV+o7HxlloDvGC1csPttx/13O2bN7N57lyC168ndONGtn30EZvXretwmw8wenRfysoWsnTpB2zbtpi8vB8YPFghNbV3q+qZNm0ajz76e2y2UYSFjcNu78WwYSv4/PP3O8LsTsGnqZgQNHdKhQNurX0SjXYXdi1bQP7WCnzuTcBDqP5/4PfOY8ELj+L3tTwDy/FyRlIGLyoKKxs+e4EHgSBbEGnBR57Lrff7+DxrMzJ3FxMqS4gr2svXOzdQ6Oqc9501LIrI3kaqi26jKv/POCvexO95gdRTRrR6H6ehs6/DkbgexZiI0TIaxTiA3mNGEtv/lA6yvpNdfDMuvJA3V6/mfJerURm3ALtUlbFjx3amKR3K0qVLEaIv0HR+w4rT+Vvee+/TVq8/aim5ubkUFxeTmZlJcHAgy9dH8+dz1fnnk7huHZEmE3VGIy+9/jojRhx5cra2tpa8ZcuYHB+PqSGAJVlVWbJ6NZVpaYSHh3eI/QcoLi5m3rzNBAWNRUoPLlcuGRlWJkwY2abIxzvuuI3rr7+ajRs3EhMTQ79+LfO7d1fCTBZirDY+qXc2Lr4G+LsQDIvspq6mNrJ9yRJ8nlsIbIxzgFOAVIp2bCB5cPu/N1Qp2VdXjVFRSLaHIISgd3AY1w0YyZydGwiWkiqp0TvYwV2Dxh7Vbb+hrIjRbidzbIHuxEggzefh4/ws5vQZ2qEufyk19q7+gfK9dixBV+L3lKH5fiJpzCAcCYdZiXwMTNZgxlx6D67KYjx1lQRHp2Bqh7RMR6NTBeqqq6/mv2+9xeSff+ZSp5MCg4HXzWZefO21445s604EohIP17vzYjK1/1deXV3Nleefz7KlS0kxm9nj8/Gn/2fvvAOjqLo+/MxsT++9AwmEhF6kCVJFLICiCCp2Rf1s2Nv7qtgQfbH3hoqKKCAoFpr03msIBNJ73b5Tvj+WlgJJIAWU5y+YnblzNrs7595zz/md555jyhNPEBYWxu+rVpGVlUV5eTnt27evV2WjqKiIcDjunAC0Gg2RQGFBQbM6KEVRWLJkB56evQgNdd9HVbuRlbWRI0cySUg4M2UDb29vBgwY0JSmthqCIHBbhx7ctW0VK1SVrqrCAlHDZq2OFxOSW9u8JsWdpVfXb8nVLAWn20sL+XDPBjwVBYeqotcb+L/UPsR5+dInJIpeQRFkWavw1OoaVK+UV17E6Bqh0mStHqfNgkVyNWsYtargMCWHdRh97kMQ3L9lWRpE9vbXCYhJPuNOuB7+YXj4111z2dS0aIjPaDTyx6pV3PnRR6wdNw7z5MksXr+eiTfe2JJmNDv9+vVDq80FTu5vU4Gn59vcdtv4Jr/fnRMnErpiBZl2O/MqK/ncZuPj//6XTz/99LiiRnR0NKmpqQ2SgNJoNNRV/eIEtM0gIXUyFRUVVFUZ8PI64QTd6h0xHDxY0Kz3Pp9o7xvItN7DKI5ux3fBEQQldOS13sPO2WyyM6XjkGHoDP/DrSl3jN8QtYWEJ9XWyzsbShw2/rdzLTNdTtbLEm8qMgPsVqZuWk7h0b5PGlEkzsu3wcW0Oo0WWw1VGycgCaA7g2hAYyjPzUHU9DnunAA02kBUpS3W0txmvXdT0eJp5nq9nokTJzJx4sSWvnWLodPpmDfvO0aNuhpV7YvLFYYozufGG8czatSoRo2VkZGB0+kkMTGxznBAUVERfy5eTLbTyW7cwpW9gPPVX80AACAASURBVFccDr6fOpWenTvTuWfPWtedjvDwcPYajZRWVRHg7c7QqbRaydXpGBQZ2aixGotGo0FVa++jyLKEXn/mNS//RIKNHoxv0/jyjPOJNr0uIeniTez7OxFVvQpRkwvCWq584g3ERhTJ2p0u9ucXobM7qB4uPMHf+ZmMQyUFeAX372gKEKHKzNm9gauSezS61iwuMJzfM9O4TaPFIAioqspSh51gv6BT1lw1FaJWg6raar+g2hA0rVPf1ljOayWJc5kBAwaQk3OQuXPnUl5ezrBhD9ChQ01pylOzf/9+bho7lsyMDPSiiEdAAJ/MmlVLk66kpIQAnQ6bw4EZuBj3sjgVWGqzUbRpE9v1et569VWWL1lCcGAgd06Zwq233XbK+Lder6fbqFFs/OMPfHJyEIByo5HOo0ZhNDbvDN3Hx4fISC1FRUcIDo4FQJJcWK3pJCae33tHF2g8giAw9O7H6HrZWDJ3rsfoGUOb3k+hNzV87+Ojjz7hpedfIFBVyXdJdAsI5c4OPTHVEAiocNjoqygsAIbhFkHm6L/jVYU1ORkc9g3g98P7KHXaaesTwJiEjsR5+XIqUvyDWWkz82JxHu0EyFVB8vJlWDMoRtQkIKYNBfv+RpG6ImrdNrps+9GZcvEMPD8EYy84qLNAkiQqKirw8/NDU0dFu7e3Nzfd1PgKa6fTyYiLL+axoiLuUlVE4BeLhTGXXsqugweryQq1adMGu1bL30AKJ2K2K0WRTl27oi8vZ9iAAdxnsbBIUcgsKuKJBx/kcFoaL06bdkobQkJCGDphAsXFxaiqSlBQ0BkrfjSWwYO78/vv68nOzkIQTAhCEf37xxAe/s9KALjACSwWC5bT9CkKjGl7tPV44/j1119558WprLQ7aI9byf+ukgI+2buJ+1OrP6Q7+Ifwff4RBsny8eQTF5CJQC+jJ9+UFrA85xAzVIVk4JeSfF4oK+K/PS4hxtOHuhAFgYGRCZQHR1Bst9FNpyfU6NEi9ZAefqHEdI8ja+urqGoHUM3oTJm0HTAQUTw/ohEXHNQZoKoqb7z6KtNffRWHw4GHycST//kP9z7wQKO/eDk5OUyd+jp//rmCsLAQHn30blRVJd5m456TYtdXAQtkma+/+opHHnvs+HGdTse0t9/mnjvv5Cm7HTuwXqtlrcnE69ddxxszZ3KlzcZzirsaKglItVho/847PPTEEwQE1Ez8P4FGo2lyjb2G4OXlxdVXD6a4uBiHw0FgYPJ5LUl0gVOTX1rOw1dfy+JVq1FVldDodvS//0WCYhqnEweQsWUVm+bOxlxSTGj7ZIaFP85n06cz1Wqj/dFzvIAPVIWoknwqXQ58dCfCfT0Dw/jD05e/K0vZCUQA2xAIMnmgarRkW6v4BfdEEOBh3Cn/8zP28H8pp1+R+OmNrbI/GNwmFf+oNphLchA1XngF9zxvnBNccFBnxDv/+x+zXnqJZRYLHYAdDgfjn34aT29vbmmERl1eXh6dO/ehouJaJOlDDh3K4IYbHmfo0E4kSrX3YRLtdnIza6oiwMQbbyQsIoJ3pkzhUGkpXbt2ZfqYMbhEkbUZGTxeY6wwoIPBwJ49e87ZNhaCIJwzRdcXaB5kRWX0lFe4sriMbyUJPfBZxj6eeuYWxr/3K0bvU4fOarLjz59Z8eXnSI6XgSQqi+Zx6xWjiPbWUFNXwwcIEEWqXM5qDkojijzZZQBfHtjBjMIsxgkiMV4+BBk9+cZmxoJASg1VhsuADytLz/RP0CJoDR74RTTe4Z8LXHBQZ8Abr7zCvKPOCdwNAz+yWrn7+ecb5aCmT3+LqqoxSNL0o0d6YbH04o8/uhIgVG/JpgDzvLyYcoq+S0OGDKHz4sVsXbwYfVkZO+125MBAIrt1Y3dWFiOVE3oSNiDd4WhwF98LXKA5WJdbgLGiipdOmkBNRmW5y8W+vxfQ5fIbGjSO7HKxaub7SI6lHFvfqEpPbFZQQubzs6aA7icpuGzHLdgTUkcNj16j4Y6kLqz1DWB+UQ6RQJ7Tjk9gGObKUopUOHnatOMU41ygaWhVB2W329HpdHXu35yryLJMVnExXWoc7wpk5Oc3aqzFi9fidP6nxtF49Po2dOrmw4gNG3jcasUEvGs0omnXjquuuuqU4wUFBTH0uusoLy9HEAR8fX3xb9OGkX/8QRerlSFAGfCgwcCgSy654KD+ISiqiktR0Itii2o9ni0ZlVV0qqMtTm+nne9yjzR4nIqCbFTVmxPBNzeKPBarYx5feHkim82MlhX2Ac+IGq5vm3rKNG9BEOgbHkuX4AjKnHY66/T46Azk2MzcVJDF54pCOLAeeEzUcEdc+zrHucDZ0yoNC1evXk3f1FR8PD0J9PbmocmTsdnqSIc8B3n88efwEjQsrnH8T6Br+8Z9UWNiInC3KjsZO05nJu9//jkTpk/n9S5deCY5md7PPssfq1bVm6ggCAL+/v74+fkhCALdunXj89mzuS8ykmCjkTiDAd2YMXwxe3ajbD1XqaysJDc3l8rKyvpP/ochqypzDu3h9pULuGHFL0xZ9wfri3Lqv/AcoNxp5/MdmfwmQc1g9gKjiYB2DW9hYfL1R5FLqF4rBbCf2NgYlq1YRvmw/kwO9OebsGDuSL2IIQ1oZe+h1RHp4X08DHhTYlcMYbEkiSJBosgYnZ7rkrrSNaDl92mbGkWRqSrKpDI/A7kFtQLro8VXUPv27WP08OG8Y7VyDZBvs/HAl19ye34+386dW+c1qqqycOFCZn38MXarlSsmTuSGG25Ar29ZMcsdO3bw/vtfYFM/4zru4SOs9AFWAA8ZjXw9fXp9Q1TjscfuYenS67FaU4FAoASt9j369u1DfHw8d0+ezN2TJ5+13aNGjeKyrCwKCgpQFIWsrCwqKyuPyyGdj8iyzKpVm9m9uwKNxh9FKSM52Zf+/bufVyvys2H2wV1k5hxijSKTCCy2W7lhzyZMnXR08q+7y6vZ5eSPnEOklxbga/RgaFRb2vo0r3RVXXxz6CDl9usxsIcr2MDL2DEBbwL7ffy5uu/w+oY4jsnbj/gegzi08S4U6X7gAJCH3vA2Ux55j5joaN544GZKtu6lPO0gGQ1fnFVDJ4rcktSVCW07YZGc2CQJq+xyayOeR4kHNbGU5XFw5UpcjmhAjyhuJK5XZ/yjW7+so8Ud1LvTp3Ofw8ExPYUo4Gu7nZjffyczM7NW2ElVVe6+5Rb+nj2bR2w2vIAP169nzsyZLFiypEUfRgsWLMDluh6YRBmh3MbTyBxEKxi5ftJohg9v+I8K3LVSr7/+NI88cgey3B5ZlkhKiuTmm8fjcDiavKfVR++8w4w336SNXk+G08mI4cP59LvvzkuZqZ0797Frl5aoqCGIooiiKOzatRU/v/107vzPkvupC4cssyjnELsUmWMdm4YBrysy72fsrdNBFdktPLdpOQMlF4+pCvsrSni1KJeb23ejfwv3fVpfVIjMg1iJ4S+msoovUHHiopwJ//2k0V1yh9/7OD8++xBFhx9FEAYBTiKiL6Gy0l7fpY3GLDl5a8da8qxVBAsC+cBN7TozKDy2ye/V3Ciyi/QVK1GVWzF6u9NJZFcxGetm4OEfgsGr5ScvJ9PiDurArl2MqtFywgPoaDBw8ODBag5q9+7dXHv55WQePswROK6Efo3FQp/Nm1mwYAGjR49uMdvdbdotR/93KVVcCoDJeAspKSmnvvA0RER04N13/8RuV/H29sbPz4/s7F3s3ZtOly5NpxLw5RdfMO+tt9httxNht2MBbv/zTx66+24+mjmzye7TUmzfnkVo6MDj4rGiKBIa2oFt25YTEuKPJEkEBQX9Y9PTK1wOPIGabqUX8KyttlL2gsw0vj+4i3HAFycdH6zIjErbxkXBkWibWXrnZLSiBneqggmZlzDzElCFqAnF0/fUpQ+nwlZRhodfT5L6P4wig2DwoG2glu9mvcyAnik0VRBOVVWmb1/NOGsVz6oqWtxJF8PTthHp6U07n8bb3pqYi7KQHG0w+pzIddTognBa+1F0cAeegWFodAa8gqNbJT29xfegOvXuzbIaem4VwE6Ho5rSgt1u59KBA+l1+DCjoVqbDi1wvdnMkt9+awmTjzNu3Dg0mh+BvScd3QnM5+qrrz7ttaWlpRw6dIjc3FyUoxl1drudvDwHwcFRREdHH2/BHhAQw4EDTdt/6cPXX+e1o23fwd1y7B27ne9nz8ZqtTbpvVoCp1NGq63+PbLZHKxcuYt5847w669FzJz5N/v2HWglC5sXf70BuyDU2sFcCsR4Vk/PXluUw7KMvSQCt9Y4vyfgo6pkW09dJNscDAkLxaB5hhM7UCoa7UtEpPRHdxqdO8nlJC9tB5k71lNVfCIpqTgrHVXthc7og8HTB1EU0Wo9UOnK3n37mszuw5YKLDYL/znqnAA6A48oMkuyDzbZfVoKWXJBzcYtqorTLHFk43YOrYUDf+ez+7e52CtLWty+Fl9B3ffww/T64gvCJYkJqko28IiHBxPGjycs7IRC7oIFC2jvdHIV8HYd4+TodAS0cJ1MTEwMH3wwg7vv7otWOwRQkeWlfPbZh6dUOVAUhZUrN7F7txlRDEFRCggJ2cvIkReh1+sRRQVZlquFKiXJiY9P0340hcXF1NwWDgR0gkBlZeV5F+Zr1y6Y9PRMQkPdkjHuv/MqQkM7ERXlLpp0uRwsXbqK4OCA87p7bl3oRA2jY5MYc3gf7ysyKcCvwLOihidrKJovPpLGa4rMZ0BNiVAHUKoqeGmbVwS4JtfGteWIcxu7yxNAGIao2YpfkMTFd596H7eiMIc1s37Cbm4DBCAwj4ReUaQOG4VWp0eg9spRwIzB4As0jQMudzqIEYRaM/sEYJ7j/Ej0OhmvoEgQfkeRqxA1bp1Bh7Ucu3k1ftFTMHi4k1Uclq0cXPszycNHt2imaIuvoGJjY1m2bh2rhg2jg4cH48PCGPn007z18cfVzsvPz6ety8VIYD8wmxONizcDM7VabrrVPR+srKwkJyeH8vLyZrd/0qQbyco6wHvvXcn7748mO/sg119/6v5O6emH2LlTJTJyEJGRKURH96G8PJ7Vq7ej0+no0CGI/Pz9x8+XZYnS0v107BjVpHZfPGgQs2uEcJYCQYGBraIWcbb06JGMh8dBsrO3U1SUxf79q5Gkw3TufKLwWKczoNfHc+hQdita2nxcGZPIJe06cZvRg3hRw9s+/jzauX+tMFO5005b4HbgeSDv6HEZeAZo6+1PkNEDWVE4Yq4gvbIMm1SXnn3Toddo+PTS7vz+2s1Me7kjP/wwlW//WoyHX90ipqqqsvHnBUjOCXgH3Yl30DV4BDxD+roq8g/sJLRtChrdJpz2E1mMlqpDmEx76dKp4RmB9dHO258dqkLNb9R3oobEwJZpQdGU6IxeRHdpg9MyA3vl3zjM67GUvIneMwi96cS2hd6jC/YKT+yVxS1qX6vUQSUnJ/PTH3+c9pz+/fvzuijyJrAQGAe8ePS1HJOJT7/+mvj4eDauXEn5rl0EiCLlsoxHYiI9Bg5sUFuJMyUoKKjBGnt79+bi79+xWqO94OA4Dh1Kw26307t3J6zWjRw6tBRR9AZK6dMnvNGdY+vjmZdf5uK//qLSYuFSSWK7IPCqycQnH3xwXtXOHMMth3Qxhw9nUlJSjKpqMZk6YzJVz0wURS2SVLPp/T8DQRAYFhHPsHpSphP9gvmxIJOpuIPTybgrhvYCkV6+PJzSmwKbheUZewk/ure1URBJiUwgNbB5Jy/dEuPp0ftyBA8fNuWcepVTVZSHuUSPZ8CJFhuiaEBrGEzWzhWEJ3ai97jhbPx5BmZrHLIsoUaW8uxzt+Hp6UlTfQO8dHrGxLZn4JH9PKvIRAJfCiJb9AZeakDq+rlISLuueAWFU5q5A0VWMHo7sZtHVHsuCIIAggFFbtmOzeeskkTXrl0ZeOmlDPv9dx6zWnkHeMlgwJqQQNbGjXh6erJv926kLVtIEkWqCguJ0GopKCtjj69vo1tMNBeKotbqAisIAqrq/vANBgOXXtqfsrIybDYbfn4dmyXc1q5dO9Zt386M117jv6tXE9uuHQsef5ye58jf6UwwGAwkJbklXCRJIi1tMVZrFR5HWyIoioLNdoTY2DataWarMzq+A0+X5OGUJK5CxROYKohcFteea+LaIysKvx/cwjinnVBZwiVJ9NNo+CozjRAPL0IboRzeXKiogKbWZEoQNCiy2/2EtunApQ+2oSwnA6tL4f7RF5McF4BqbdoauTFx7Yn28uXT7HQsLicdgyJ4MaoNHi0cJm1KTm5CWJq5l0Nr1qCqHREE97NLcmSj1Rdi8mtZabRz1kEBfDl7Np9/9hlvffwxLpeLsTfeyOR77z2emXVk61Z8s7KQLBZiPD1xSRI2s5mNv/1Gpx49zomVQfv2YSxdeghPz27H7SkpySEmxqNa6wp/f/9mb6UeGxvL/95/v1nv0VpotVqGDk1h0aI1lJXFIIp6JCmbzp09//Uq6KEmT17pOYQFR/Zzd3kxAQYTD8Umkerv3sPNsZnxs1nws1sJBTwFgQrZRYqisLUol0vPQLi1PuJjIfaqYcf/f7rVE4BPUDgmnyrs5jSMXu6MM1WVcdlXEJ1ywj6tTk9wXBJmp8S3W/LQ6nR09weh+3ACgfK0pklk6BEUTo+g1vteHfv7Cd2rl7b8+ee2sx7bLyoR/+jFlGW/gyD2ALUMQVxD2wE9WjyT75x2UBqNhjvuvJM77ryzztcLcnOJLC8n4aQHkKfRyOIDB7DZbOfExn+7dm3IylpPevpqNJpQFKUKH59iBgw4P/qxnE9ER0cxYYIfmZnZOBw2IiLaExISck5MVFqbYKMHtyZ1rfM1pyzjsltJFAR8j672vYBYRWZleTE0sYM69nDV9bsCQatnc5nAV+szQQAvfd2PJEEU6TlmFGtmfUpVcVcgANQtxHQyEZ7Uudb5XnotZqfkHrd3DKqipXv34bQB+GFhk76f1uDY329zhRZBFPhqQyarlu1CFCDa/+yee6KoIaHvMKoKjlBZuBWd0YB/1DD0Hg0X720qzmkHVR8ab28sNZS68y0WggMDz5nMNK1Wy7BhfenUqYCSkjI8PQOIjGxY6/V/Anv27OGLL76hoqKKMWMuY8SIEbVCno3FYrGwePFiVFVl6NCh1RQxvLy8SE6+oI3WGIKNHsxVFKSTZLRkVWWXIKAo8mmubBzxR+tY21x3OUL34ScervU4p2MERMUz7L7byD+wC6e1kICofgREJZxyAnLcSW3IZFKvGDZXQI/el3M+BnwlWWbBmq38tSmNsABPJkV1pqRCiyBq+GpDJmuWn5lzspTmUpmfgck3GN+Idsf/loIg4hMWj09Y6+6rndcOKqlLF3Zv2YK+uJhAnY5yl4t8k4mgyMhm7/zaGARBICwsrFoa/b+BTz/9gvvvfwKX6zYkKZbvvnuCQYO+ZN68WWesALJgwQKuv/4WNJqugIgk3cbXX3/K2LFj6r320KFDfPDBpxw8mM3gwb25+eZJ57XcU1PhpdOjN3nyqt3KEEHAE5U1Kth1BiKaaP/p2KpJazQg9r6cTSUKMzfmgMBx59Tw8JQRMJKxtwr2bj/tmcOHu2Wdv9qQCSoIvWPo3vvys3ovLY3dbmfEFRPZs8+JxTIene4wb827nIH3PUtM94tZs3wX0DjnpCgyO375mKL07Qhif1B/weRnoOf1U+pdKamqQtHBreTt2YwoikSm9iEgtulEBU5GUFW1/rOO0qNHD3XTpk3NYsiZUFVVxYpvviHMZkMjy5g8PHCqKuUJCQy47LJWs0uSJD799DM+++xHFEXh5pvHctddd7a4dmBrUl5eTnh4PHb7OtxtEgEceHr2YebMZxg6dChWqxVfX98Gqz0UFhYSH98Bq3URbs0EgC2YTMNIT99JREREnddlZGTwyy+/8OSTLyBJt+JyJePhsYDg4H1s2rSCoKC6U5vrQhCEzaqq9mjwBXXQ1sdffaPH4LMZoslZl5+JJecQ8YK7nCNCq2O1rBAR155k/7OrN6zpnGqG9E52TDUfsrbyQg6tW0RZVgYe/kEk9BmGX2TDNOKyyk4UoA8f3gWzUwIVJvU+v1T8f5r5GR+8thSH/Tfg2MRuNaL+KhJveJlA0QIIeASEN3iPKGP9QtJXHkKRfsddqKsiiFMIjNtI92sfqPMaRZGpyDvIwZXzKMuuQpHuBeyIuneI6dabpEuubfB7+uPVcQ36HZ3XKyhvb296jB3LzmXLEMvLcQG+7drRs1+/VrNJVVVGj57AsmUFWK1TAJF9+95i7tw/WLLkl3/NfsiyZcvQ6/tgt5/8MDFgsdzCp29/jE9BAd6CQJkgENmzJyldu9b7t5kzZw6qehknnBNAN1R1DLNnz+bBBx+sdn55eTljx97IunUbsNtDUFUJ8AVuxmq9hby8u3j55em8+earTfSuz196hkSxVpHZUJyHr6qyXhBpHxVPh1PUJTWU0+03HZv5Q92zf0tpHuu+eh7JeRuoj2Ep2U7J4RdIvWISYUm96733sTGzyqzHnWDfQSnu1dR5xC+zFuKwP8IJ5wTQD1XxRbf0PUJFLSpwxOiBqe/VeAXVX0OZtWU1ivQFJ1QkBFTlBUqOhCE5rGgN1T+P4ozt7Jj/MbLkhyJZjl7XF+iG4rqZzM1JRHUZgKd/0yaOnNcOCiA0NJSQ667DbDaj1WpbXXtt7dq1LF++Bat1N+AWe7VaR7BxY1eWLFnC0KFDa10zZ84cnnlmGpmZB0hMTGHatGcbLTx7pkiSxMKFC9m7dy/JycmMGjWq3pYeDcEdYq1d2W9gHx0dFoZERiIIApIss37NGg77+RGfkHDaMc1mM5JUW+vM5QrAbK59rxtuuIvVqyNxOjNxfxY5wAigLTAep/N2fv75tgsOCnc32f4R8dhDo7FILnx0+rNW6D7ZOVXbzD+6XwKnD0sdWDEPyfkgqM8cPTIARerE3j8nEJrY83gK9DEkh40Df88hd886VEUiNLEniZeMI9r/RMjqZKfY1EjWcioPbUKRnXjHdMHgX/eKvrGYLU5qK2FUECvnMZlEYo8WZufYLXy88nuMo+5DW097edllp7qAHLidjgZZcqE9SafaXlXK1p/eRZF+Bi7Bvcb+EbgcSAeCgKsoObQdz+4XHFQtBEHA29u7tc0AYOXKlTgcV3HMObnRYTaPYcWKlbUc1MyZ3zB58rNYre8Dvdi+fRmjR09i7twvGTFiRLPaWlhYyLC+ffEqLKS/1co0Dw+eCwlh8dq1Z91uffDgwYjircAfuJ0CQBaRuu+48YpHjq+WtBoNSf7+7Nm1q14HNXLkSJ5//lJcrucBv6NHKzEYvueyy+ZVO7ekpITFi//A6czmxGcRCUwF3gLGAyXnzPfmXMGo0WLUnP1j4eSwnqDVH3dOjdnML8vcD2rNsogBSA4bDnM5Ru8TD1hVVdn43XSqijqiyisBPXl7/kdp5ov0v/NlNFp3eP1sM9xORf7etRz59V2uAHwUhdmb5hLa4zLiBk0867F1vfqz69dXkF1X4I4AAEylj6gS43vidxpp9KRrZQnbCzLwj+5Q51jHCGnXiZydH6Eqb510dA4m3zD0Hj7Vzs3dvRJVHYvbOYF74/Ba3LLDvwDjEYRiNPqmD522SsPCfzJhYWEYjbVrLTw8DhIeXj1JQlVVnnjieazWb4CRuNXxrsFme5eHHnqu2W194oEHGHzkCKuqqnhNlllVVcXgzEwev//+sx7bYDCwYMGP+PjchLf3CDw9J6DXp9AuIYJvv5nLtGlvse+oiKdBq0Wy198WITU1lVtvvR5Pz164OwfNwNOzNxMmXEW3bt2qnVteXo5W64s7YfpkYoFioAoPjxe4996GKYJcoPFojQb0YbXDTQ11EjoPP6Dmb6kEVGetEFRp5m4sJU5U+RsgEYhDVd7CYY4hf++aM7K/objsZg78+i6rJSffSU4+UiTSJCdlmxdRnlNTzrfxhCZdRERKW0RtAqJuAhp9X0TthwTojRQd2E5FXsZR0VfwVVVkyVnvmG0HjEbv8ROidizwOYLm/9Do7iZ11E21Qu1OSxWqXNfk8dhv6W9gBSHtmr7o/1/loMrLy1m+fPnxB2NzcPXVV6PRrAe+w70UVoGf0WgWc9111TX7HA4HBQVHcMdyT2Ywe/duR5JOLysye/aPJCR0RqPREReXwtdff9soW2f//DNPShLHvo4C8ITLxeyff27UOKeif//+5Ocf5vPP7+DVV/sSFBTKlkMBbNnbl1WrYnnmmVdZsmQZWWVlBLdt26Ax3377debP/4Cbbkrjxhv38vPPb/Hxx7XlhOPi4jCZBKDmw2kmOp0LozGea6/twp133nH2b/RfSLalip1lRZhd9T8Mz5SEiy5B1D0CHNszMiNq7iG0fd9aISxz4RFkaSA1H2mqMoqsrX+f9j6S08bu379i8Ru38+e0G9g0ewbWsvzTXnMyRQe30k8UObkaKxi4R3JSvGdlg8c5FYIgkDziRvre9l86DPUjumsoqqpjtS0Es7ULVYUiBfs2YnPY2CwIeAZG1jumwcuf/ne8TOIgf0ISZxHXM59+t79UZwJKYFwyGt13VO99bAbmIGo/Q6MfQ9er70dnbHrFkX9EiK8hTHvpJV6ZOpUUg4EMl4vEjh35fuFCQkLq7jx6pnh5ebF06a+MHXsjxcVPASJ+flrmzPmFgIDqMV+DwYDB4IPNth3octIrGxAEH/766y9GjhxZ532+//4Hbr31EWy2t4CRHDmynrvvvh1FUZg06cYG2apSe4ai4YQob1NgMpm45pprePbZ5ykp6YvD9Qa7mU9bXBiccUz9cBqPff4RlyQ3rMmgIAgMGTKEIUOGnPY8jUbDBx+8waRJV2OzPYaqdkCv/wWTaR6vvPIsl156KfHx56d2WmtS4XQwY8caci2VxAkCe1SFMTFJ514QPAAAIABJREFUjI0/fUjpTAjveDHW8lIy1qUgijEocjZBCV3peOnttc71CAgHdT7ub+/JK4B1mAtP3UL3WGiwsiAKlMVAW0oyPmDdzBcZcNdr6IwNK0Ooa6YvqkAjsqTrw9M/HJNvCMvfeQBVXsg+dvIdc+mjtgfJyeacA1T2v4Zgn4YltmgNHsT2GElsPbl0QQld8AlfTEXuYBTpAcCBRjcNn7BY4noPJzA2BY2uaZurHrexWUY9x5g/fz6fvfIKO+12oux2JODprVuZdPXVLFp59jOcmnTr1o2MjF3s2bMHVVXp2LFjnRlqgiCQmtqRDRsm4l5xpQLrgHvQapPJza3ZHMFNcXExDzwwDZvtGdxztVVAV6zWL3j66Tsa7KCuvuIKXp8/n9dOWqm9rtVyzZVXNvId18+CBUtxOJ4FAsnjOgo4iIEyNNpv8IuPb5a6tWuuuZqYmGimT3+fjIzfGDy4Nw89tOVfV4/WlLy3az2DzOVMO9oPKRu4JDONCC8fLgquf+beGARBoG3/McT1vBRLaS5G74BTdngNiu+MuzHP48DTgA54H9iMLJlRVbXWb1BVVdJXzqcyXwV1KO6eCTGgPoTs2kHO9uXENaBmKiihCysVmT24hXgByoD3dXpikptWu85Wlo8i6YG+2OnDGnqxnZWotENW5zAwdVCT3g/cRbs9rnuInB3Lyd31GoJGQ3SXQYR16FsrUaWp+Vc4qE/efJPnLBaORcO1wAuSRNSmTWRlZREd3fTtrgVBoGPH+ovX7rvvdm677SVcrstwx3MjgSfQav/DRRfVlkNyOp0sXLiZoqIw4GbcCQDFuJuQDCInZz+KojRIreG1d95h8MaNbCotpb/ZzCovL3IDAlj2zjuNeasNIigoAHcWHYAnCp2w4cKkVDRru49evXoxe3av+k+8QL2UOGwcqCplxUnN+qKAFxSZGVnpTe6gjqE1mPANP73+gyBq8AyKxlK8DggDFGAY8AhewZ/UOUEsObKLgjQd8AowCncDkm+BH1CkQVQWNCzUrTd50/bSu+j9+8dcqyr4KgozNVoCOg3BP6ppVU20Rk9UpRKwASagKxa6Aj/h47Ws2cpYRI2O6K7DiO46rP6Tm/K+LXq3VqK0uJiaW7UGIFino7S0tDVMOs61115L27beGAwDcDcWeRcPj6+54orhdTq43NxcbLYQAgJigDnAEdxpnsHAAiIiEhssJRQeHs62tDTu+OQTeO457vjkE7alpTXLCuOhh27H03MqJ/YTZDSaF0hJSSahnuy9C5wbVLmcBAgiNYM50YDZ6WgNk6qRPPx6RO1u4GVgJTAcUfs87YfW3e26MO0weo+rgBIgA/cey9XAZgTNMrxDG/47CE8ZSLc7ZrD64gks6H8tiTe+RMKwW872LdXC4OmHX1QHBPFJTuwJ5SPqniau58Amv19r869YQV0yahSz0tMZ6DyxobsVKBPFam3mWwODwcC6dUt4/fX/8cMPT2A0Grn77oncdVfdArlms5l3351JRUUV7o4+fwJtgPYYjY8ydeoLjb7/+PHjz/ZtoKoqf/31F999NxeNRsNNN13LxRdffPz1UaNG8dRTe3jxxc4YDJ1wuQ7Trl008+f/WG0ci8XC9u37ycgowWTS0aVLLPHxcf+aAudzmSgPbyoEgc1A95OOfy2IJAe1ftg0ICaZXhMf4+DqeVQVvYdXUCRt+z+Kb3jdCTjWshLKs7NBDQF2AMuBnkAGGu06IjtNa9T9Tb7BxPW+4izfBbjsFnJ2LKciNxuv4GCiOl9SLbTZ+ao72DLnHaoKoxE1CSjSLmJ7jCSsQ/Vkq7KcAxTsS8NhtuEd7E94ciomv6bdc29uzmupo4ZSVFRE3y5d6FdaytV2O+mCwOsmE69/+CETb2zYfs25wr33PsTHH2cjSV/j1iSTgFfR6eYyY8Yd3HPP3S1uk6qq3H77ffzwwxIsljsRBAmT6QPuu+9GXnutusMsLy9n8+bNhIaGkpKSUu01m83Gzz+vxGJJICAgEqfTRmnpXvr396NLl+bR+mos/1Spo4ayMj+Lmfu38Lgikwj8KIgs1el5qedg/E7KrIuPdYvC6sOikGK7uBUkjtZBNVctUmNQJBeL/3c/qvwycCyTsxyYgc7jB3pNnIxXYNMU2jYGe2UJa798HsnZD0UagahZj6j5mZ4Tn8QnNK7auebibBzmMrxD49CbqtfzFWfs5vCGXLT6MWh0Ybjs+0GYS4ehAzH5nl2NY1PQUKmjf0WILzg4mPU7d9LuySd5v08fdowbx89LlpzWOUmSxMsvv0ZYWFs8PPwZPnwsu3fvbkGr6+bbb79Hkm4ANgBZuBt4pyIIGUyefFer2LR+/Xq+//43LJaNwMOo6mNYrRt5++0PSUurXgfi5+fHkCFDajkngPT0w1RVRRIe3haDwYS3dwCRkb3YuDELh6P1Q0gXgAFh0TzaZQB/hkQy1ScAJTaJV3oNqeacanIoI4Mn7ryLLycNZv9Xj5O2fDaK3Lwt5eujOGM7ghiPWwnhW9wxlZXAeoISolrFOQHsX/YjTtskFOlH4HYU+RMk5yvsXvRNrXO9gqIIjEut5ZwURSZnxx70HreiMyUian0wePVEVS6nYH/rP8Maw78ixAcQEBDA0889B881rAD2zjsf4Icf9mC1zgaiWbx4Fn36DGbnzg3ExsY2r7GnwWarAvoDDtzOCSAFl6uizkyllmDhwt+w2SYAJ/9QglDVsSxatIjExMQGjZOXV4GnZ/X9KK1Whyx7YzabMRiaJ5X1Ao0j0TeARN/6tfAAikrLGXDNzZRXTEZV3gOKOLLpMczFH9Htmvua19DTIDltCETgzvj7G1iLey+3H7C61ewqOrgN1M9qHL2ZqoL/Q3Y5GpTOLTmsSE4jRu/qKyWdsR3m4j+a0Nrm51+xgmos+fn5zJo1C6t1HtANCEZVH8Buv4k33ni3VW0bNGgEgvAFEIF7J6A78Cd9+w4/6z5LZ4qXlyc6XXmt4xpNOZ6eDS/eCwryxGarqHZMURTA3Ooaixc4Mz6eswib7VJU5Rnc2XWpKNJcSg7vxlJadxlFSxAYm4KqLMNdcHoFcA8wBo3uJ0ITU1vNLo3WgDvUeDJmEESEBv6+tXojosaGIlVvdS85czD6tH54tTFccFB1sG/fPozGVE7oXrlxuQazYcOO1jHqKG+//TK+vm9iNN4FfIvBcB8+Ps/zwQeN29BtSsaPvw6N5nvgZCHOdcBfjBlTf5+mY7RrF4coHqKsLB9VVXG5HGRnbyM1NeicaD55gcazYVcmdnvN/TYjgtibqsLWUxU3ePmT0PdKRF1v4HXgUzS6vvhGGghp272+y5uNyE79ELVPAccSuhQE8UlC2vVF1DSsyamo0RHWPh6H+TtkqdT9W3JkoMrzCGvfuklhjeVfE+JrDAkJCTgcewArJ+ToQavdQGpqw8JVzUVSUhL79m3lgw8+YdOmhXTpksQ992w5ZS+kliAuLo5PPnmXO+4YgFbbD0FwIcubmT37awIDAxs8jre3N6NHd2fNmj3k5GzBYBC46KLIBidIuFwucnNzcTgcBAQENKrP0wWah9S24SxZvx6n82TRVAlV3YpnwCWnvK4laNPvKgJiEsnatgTZ6SSsQ39C21+EcJYq7mdD2/5jqCp8j9LMWAShP7AFjwAPOo6c0qhxwjp0B3ELBXun4bIKGLy1xPXs1KBWHOBO1qgqykTUaPEJS2gWGaOG8K/I4jsTxo69gUWLbNjtb+EOTczBw+P/2Lx5Je3bt05LcVmWycnJoaysCj8/LyIjI5ukNUZTUVFRwZ9//olWq2X48OGNCu/VxOVyodFoGhy2LC8vZ+HCDVRVBQIeqGo+qale9OvXvUlDn//2LL6GciyLr1Aw0OmaBzGbXwFuBEoRNI/iF7GPXhMfbzX7HJZyynPSUSQZn7BoPPzDz6lShqrCI1QVHsHDP6xaK/bGoqoKiuRC1OobNIaqquTt2UDe7gJUtSdgQ9RsIaFvV/wiTl8s3Rj+FQ0Lm5NZsz7l4Yef4ssvU3A4LHTs2IsPPpjXas7Jbrfz669rKCjwRqsNQJKKCAo6wOWX9z1nwl++vr6MGzeuScbS6RoWzjjGsmVbkeVUoqLc/WgUJYnt29cTHZ1JXFxck9h0gcYTFRrMnwu/59b7n2f/zskg6AjvMIAOw85eMf9MKcvaT8a6XahKP1TBRO7O9YQkHiSqS/9zxkl5h8TiHXL2yViCIDZKJ89alkfe7lL0Xo8jiu59X8nZj4x1b9Hpiqhm09w7FRcc1CkwGo28//6bvPvudFwuV6tnkG3btpeiogiio084yLy8A2zatJuLL256mfvziaqqKgoKFCIjTzRLE0URH58E9u8/dMFBtTJdOqXyydyf+HzNQdat2EtMYMMEWJsDyWnn8IbtaI2PoNG5Q8Cq0peCtLfwj8rGK7jpZc/OJ8pzjgD9jzsnAK0+HLs9EXNxdr2SU03NP9pBqapKWVkZXl5e6PX6MxpDFMVmc047d+4kPz+f7t2711I6r8nevfmEhFQPC4WExLNv3x8MGNC86eWyLLNnTxo7d2bjcskkJYXSuXP7C5l1/yJcioxdlvHS6s74u6bR6hqcidYYZJeD8twDaLR6fCPanlbA1FKSg6Ikoded2J8URAOC2Jfy3K3N7qBslcXk7tpGZX4xOqORsPYJBMannjMrNzd1bPsITdnjoOH8Yx3U/PnzeeK++8gtLEQQRW65+WZenTGj1VdCAHl5eYwcOY709Gy02ngcjm08+eTjPPfcE6e8RqMRURS52jFFUdBohDpVmletWsXWrVuJj49n5MiRZ7VXtWrVZnbtEgkO7oPBoGX79iNkZq5mzJiBjQ7FNQfe3t6EhoqUlOQSeLTAUlEUKisPMnBg8wiY/ltwyjLfHNjOsoJMVBWC9QYmJHahZ1DTtvY+U/L2rmX3b58jCG1RqUSrt9Jt3AO1VBeO4XZedRQJqy5ETW3HJjlsFB7YiMthITCuE14N6LV0KhyWcvYvWY4iX4XOlIrsKuXwhvk4bRuJ6HhuCBr7RcaRv3cVitwLUePeOpCcuYiaA3gFXdXi9vwj08zXrFnD3RMm8G52NuVOJ7vtdg5+9RX333FuNKcbM+ZGdu8ehMVykIqKZdjtu5g27Uvmz59/ymtSUyMpKEjj5KSW/Pz9pKRUz96zWq307z+CkSPv4LHH0pg48VXatu1Ednb2GdlaUVHBnj2VREd3x2TyQq83EhGRRElJAJmZWWc0ZnNwySVd0el2kZ29iezsPeTkLKdLF32zKNX/m/h032bUgkz2KwqVqsJHDhsf7d5AWkXriiyDezW069cvkV1LkZwbkZ37cJhfZdP300+pVOEVFIVWdwjJcfj4MUWuAlbgF1W9SLwsay/L33uAPX8eIG0prP1iKnv+/JrGJJadTFH6LmRpMAav3ogaD7SGKAzek8jfewTZdW4opXj4hxGRGozT8iq2il+wV8xGcb1NQp8eLb7/BP/QFdSMqVN5zmrlWEu7SOArm42EH3/klRkz6g2nNSeHDx9mx46dSNIi3O0B3RZaLM8wY8ZnXHVV3bOUlJQkioo2kJ7+N6IYgKKUERcn0rVr9Yr+F198lS1bfLDbdwMaHA6wWp9n0qR7WbLk1A4Q3PVfTz/9EitXriY0NJynnrqPAQMGIAgBtVZpen0gJSXltGnikHRlZSXZ2bnIskJkZFiDPys/Pz+uu27w8TTzwMAurfo5/xOocDpYV5RLpqocrwgcBjyryMzL3E9iap/WNI/s7StQlduArkePCMBEFPldijN21FnPJGp1tOnfl/SVH2KvTAJMIOwgultbPPxOtH1RZImtP72D7JwFXHr0aDm5O3sRnLCF4NPUSqmqSt7ulRxatxSntYyA6Pa0G3gV5hIzGn31H4yo8UJVg3BaKppUyFVRZKoKDmMpKUTn4YF/ZDu0hvqTqQRBILxDT/yj2mAuykTQaPENu7xB1zYH/0gHlZGeTs2qAX8gUq8nJyenVR9cZWVlaLXBuBuqnUwkxcWnnpXqdDqGD+9H9+4lVFVV4eUVQWBgYC3H8dVX32O3z+aE8wNZfpSVK4Mxm814edW9QZ2enk6vXgMxmx9GVZ+lqOgAt9/+KFOmjCcsrH8tGSWnsxw/v6atjUhPP8TixelADIKgRZa30qdPKF271tbtqwutVktMTEyT2vRvpthhJUoU8ZWVase7A59Yza1j1Ek4bVZUpY66HjUSl+3U9nkFRZF6xWiqCo+gSC68gkfU0rMrz96PokRywjkB+CG7HiZ7x+zTOqiDq+eTsX4Lims60JaCtB8pyXie2F5XojizwHCik7Oq2BEoRlfj/meDLDk5uOovKgsCEcTOoOaTs+NXEgcOwMO/YarzRu8AjN6tP8E7b0N8GRkZPPbgg4wdOpT/PP00eXl5x1/r0qsXf2mqF9tlArkuV6v3HnJ31y3GLU55AoPhG6688vRtzAECAwOJi4sjKCiozo1Vl8sFnJwQYgVmIUky33///SlFV198cTpW62RU9XEgERiF1bqQN954i+hoyMnZicvlRFEUCgsP4+NTQGxs04XP7HY7S5emERQ0gMjI9kREJBIefjFr1xa0es+ufzKqqrKxOI+3d67lfzvWsLIgC/loCCvC5EW2qhxvM3mM34FYn7o727YkwW2S0ei+xt1o8BjFqMoSAmKST3UZABqtHr+IdgTEJNdyTgCKIoFac/5+APiTqsJDVOYfqnNcyWEjY90CFNfvwGVAIqhPI7nuxFJyBEH8A6d1F6qqIEul2Cu/JbhdOFpD0yUcFR/aSWV+AkafyRi9+2P0uQZVmcjhjevPODzZWpyXDmrDhg307tQJ4f33mbBkCcVvvEH35GQOHDgAwCPPPsvbJhMzBIFMYDFwhYcHUx599KyKR5sCvV7Pe++9iYfHZQjCdGAuJtNEQkJWM2XKA2c9flhYCDAddybOESAFmIOqPsNDD/1AUlJXCgoKal23Zs0mZHlkjaMJaDShREb60aOHSlnZYvLzFxEXl8+VV/Zp0oSTwsJCFCUUw0k/VK1Wh0YTTW5ubXsv0DR8mbaNObs3cG1xHpNK8lm6bwtv71yLqqqYtDouj2rHZaKGJbgneW8A72u0XB7bOvWAJxOS2AvvEAVRdwnwPfAxGl1vYnoMO+uWEj6hCciu3cAxYYKPgL5AOLbycaz/5n/sX/p9reusZXkIYiRQYyWvjqKqsIDES/pg8p2LvfJRFNdrRHaWiOzUtKHS0iP56Ez9qk1gdaaO2MoFXLbK01x57nFehvim3Hknb5jNHGuWcY3DQbTTyXNTpvDdL7+QlJTEkjVreOHxx5m2Zg0RISE8+MQT3HxL03e4PBNuuGECSUntmDHjY7Ky/uayywZw993v4efnd1bjyrJMevo+wAIMAcqAScB/ADCbn8Fuf5iHH36ab7/9tNq1CQmxpKfvAE7e06rA5cojNjaWwMBAevTohKqqdSozKIrCr7/+yty5i/D29uTWW2+gc+fODbZdEARUVa51XFVlNHVkV13g7MmyVLImP5MDinx8j+l6RSa1rIhd5cWk+gczLiEZf6MH92alUepy0sEnkP+2TSHco/VqmY4hihp6TniE3F0ryNvzDhqdjuiu1xGU0LX+i+uhOGMbgiYRVR6BW0z2J2A74I7AKNKTZG5JJaxDj2oNEQ3eAahyDlBFdXX/7Xj4B+EZEEH7IREoiowgiHVGQZzWCrK3/425qADfiCgiUi5ulNSQqBFR1ZpJIgogtaqM05lw3jkou93Ohl27WFrj+CRVJWXJkuP/T01N5cfffmtZ4xpBz549+fbbpi2wdTqduFx23L2ifgZuAVZUO0eSHmb+/M5AdQf11FP3s2rVBKzWZNwzxSKMxnsYPXrscT09Qaid0g5ux3jlleNZsSIds3kSGk0pn3wygjfeeIHJk+vuDFyTsLAwjMbdmM3leHm5HbXDYUUQsoiKat3N+H8q28sKuQq1miSyEZioyGwrySfVPxhBEBgWGc+wyPhTDdOqiBodUZ2HENW5/vB4Y3BZKxHojcovwIPASI45JzcBKNIt5O/bWN1BefoR3KY7RYduQZE+BAKBlYja54m/6ESERDyFozAXZ7H+65dQ5JEo0hgK0v7i0Jonuejm5zD5NExbMighmox1S9Aa4hAE916307wW71ATOmPrTywaw3k3NdVqtRi0WkpqHM8H/LybbqPxfMRkMtGuXWdgETABd6KEVOMsF6JYe14ycOBAvvhiBiEhN2IyhWE0JjJ+fAhffPFevfedP38+f/+dgdm8HngQWX4Bm201Dz/8WIP3j3Q6HSNHdsFuX0d29kZycrZQVraCYcMS8f6Xf67NhadWR14dRa25goin7swK2/8p+Mckg7gQd4+oy6nzUSm4EMTaE7bUK24jNMmMqIlH1Aah95xA6uWT8I9Kqve+uxd9i+T4D4r0NXA3ivQTTtst7F/yY4NtD4hNJritFUfVy9gr52CvfBeD1+/E9exb/8XnGOfdCkqr1XLDhAk8MmsWnzsc6HEHtB43mbh18uTWNq/V+eCDaYwaNQ67PQ1FuRiYintPSgBU9PqXGT/+WgoKCjhwIBuXS6FNm1CioqK49tpxXHPN1eTn5+Pn59dgjb8ff/wVi+UWqidntEGn68/SpUu55pprGjROWFgYN9wwhIKCAhRFISQkGaPx1J1aL3B29A6KYGbadv4ARhw9tgGYLQi8Efrvrh/zCY0npG0yhekDUFyTcU/6dgLHekXlIGq+IKTtgxSmb6UyvwSDp4Gg+CRMfiF0uuJ2pBE3IDmsGLz8TqtucQxFdlGeu50TLeiPot5D8aH6J4rHEASR2B6DCEksxlZeiM4YiVdQ7/MuvAfnoYMCmPb229yQlUXs6tV00evZ6HBw5ZVX8thTT7W2aa3OoEGDWLduKdOmvcO2bRaOHPkORVmFy9UXvX45cXEGJk16izlz9mAytUEUtezZc4Tk5FwGDeqNKIqNbt3h5eWBIFRSM0FIECobLWSr0+mIimpYS4ALnB0eWh2PdOrLTTvXEqGq6IE04N7kHgQZzw0B4tak05V3kLd7JVnbPsVhDsJe2QdBvBxUDxDmEn/RZeTs3IWtIgWNvi+VecUUpv9Fm34p+EW0Ras3otU3fILl3pPSoqpm3MHWY1QiaBqfkGTyCWpwWPBc5bx0UF5eXsz76y/2799Peno6HTt2vCAIehKpqal8/fXHAEiSxKJFi0hLSyMl5VX69evHt9+uJiJiMFqtOz4dEBDOnj2raN8+n/DwxkvY3HbbRGbNGofVehPusmiARYjiAYYMadq9gQs0LR39gni/3yj2VpQgqwqP+Aah15x/M+3mQBBEIlIGEpEyEHBLFRWmbUCRZYLbTqXy/9m77/AqqryB49+Z22+Sm947hIQuELrSRCmKfdXXDmvXXXXd911dXXtZXde2rt1V194bYqE3ASGRXpNAQnqvt9+Zef8IhISakITcwPk8zz7Pcpg55wRz72/mlN8p20tdcSSW4Itb7vF5+lCQ9TK2WalHnGc6YnuyjuiMMyjfeR+a+hrNw4o+ZN39xA85owt/st6jVwao/TIyMsjIOPa47qlMr9dz3nnntfy5sLAQTYtsCU7QvPjBaIyntLTquALU2LFjeeCBu3jkkcHo9WchSTXI8lbmzfvKL3IfCkenl2WGhHZuWfapwBQQQuLwaS1/3pv9G3rztDbX6I3xuBpteJrqMNvaf1jnfgOnX4W95lkcNenAKDRtJcFxcfSb2HPHk/SkXh2ghI4zGAxomuuQclV1YbEc/8T4vff+mdmzr2LRokUEBgYyffp0MX8knNQMJgPO+sY2ZZrmA60JnfH4HswM5kDGzX6QuuJdOGpLCYr8I7aYnk0u0JNEgDrFREVFERq6merqYsL3ZWZuaqpDpyskKantMEJDQwMlJc0ZOuLiYrHZbEetOyYmhquuuuqo1xzO1q1bKSgoYNiwYT16dL0gdERkWl9qi39A9aUg621omoq7YQEhCYFtlnPvz4vnqK/CZA0iOLbvUROvSpJEaEJGu1b9tab6vNQW7YB998u6nj9poLNEgDrFyLLMOeeMYcGCLIqLcwAdAQEOZs06rU2WjR07cli6NB9oXrCgaWs488w+ZGSkHbbe41FbW8usWZezYcN2DIaBuN3ruPbaa3j11ee79Jh2QegOtphUkkbUUbzpSVRnImhV2GL1JI+c3HKNz+Mid8V8mqqjgQFIFGKwziVj8lRMgV2XLqpy93o2ffsampYKqEhSIcMuvp3w5PblsfRXIkD5ieLiYrKz86ipcRAXF8yIEf2IiOieFTg2m42LL55CfX09iqIQGhraJiA0NTWxdOkeoqImYdj3pOfxpLJ06TISEmK7LF3U7Nm3k5WVhsfzA82/inV88MFMhg59ndtvF1sGhI5TvG7Kd66nancxmqYSnhpPTP8RHVpN1xFR/YYTnjIQZ0MVBlP/Q4JORc4GmqqGYLZd3LLJ3dW4ir3rl9FvwvTDVdlhbnsdG756GdX3PbB/FGQR67+4hEm3v9DrNue2Jh5T/UBBwV6++WYnDQ0DsNnOpKQkia+++o3q6ubtyJs3b+bdd99lyZIlqKp6jNraR5IkQkJCCA8PP+RtpaysDFWNawlOAEajGVWNo6ysrEvab2ho4KefvsfjeYoDz0khOByP869/vd0lbQinFk1TyftlISVbIkH6E5Lu/yjbnkDu8p/RVAVV8VKRm03x5mU46yu7rF2dwURgePxh34iq80sxWie0ycBiChxNQ2kdis/TJe2XbfsFOJ8DwQmaU51NpWzHmi5po6eIN6gepmkaq1fvIixsZEuKn/DweKqqNNat286rr77MwoUrkKQpSNImYmNlli374bhW27VXc0qjwwVCtcuOprbb7ciykbb5ygBiqa+v7ZI2DkdRFLxeLyaTyc+O2RY6q6llsB59AAAgAElEQVSqiMaKQMzBv2v5b2u2XYC9ppjS7avZsfATVCUVtAQ09QOSMs8ifcpl3fp7IEkSqnZQNhdNBTQkuqZdr8uB6jt0IYWqxONzdd9nyed2Islytx5kKN6gepjP56OuztsSnPaz2SJ4++3PWbCgGocjD7v9fZqaNrBnz7lcc80t3dqn2NhYZLkEl8veUuZy2dHrS7ssMMbExBAdHQO0zZeo073HOedMO/xNnaAoCllZm3jnnfm8884yvvxySZe9DQr+wd1YA1pGm4DTnIS4P9vnf4DX+SyKZzWK93NUJZfC9eupyvutW/sU0Scer2MRmnbggc9tX05oYgSyvmsWMYSnDkVn+BhwtiptQpK/IDx1aJe00ZqzroIdi79n47fz2PDNt+xevQhvq++KriTeoHqYXq8nMFCH09mExXJgrNhur2PZsqU4na8A+59QJHy++1ixIoa6urpOZz8/EqvVyowZA5k/fwU+X/MpowZDOdOnD8Ji6ZpzayRJ4j//eZHzz78Ct/sWFGUoZvOPBAUt5NFHf+mSNlpbt24T2dkqsbFTMRiMNDRU8+232Vx2mYnQ0J4/20joPKPVBlL+IeU+91ZUJYDm/JT7haN4/0Lhho+OevhgZ0X1G0ZT1WLqSp5Boj9QiDm4nMThZ3dZGyHx6UT06UvV7rEo3jsBBZ3hBaIzhmKL7tokv16XnZ1Ll6Ep/4MpaAhoXmoKl+JxLCLjzPO6/G1UBKgeJkkSo0b1YcGC9URGDsNqDaKxsYbGxi1o2sEp+wFMSJIOj6drxq+PJDk5iWuuiWo5Oyo6ekCX7WtqamqiqamJsWPHkp29nBdffI2dOz9k8uSR3HZbdpcvDnG73WzcWE58/FnodM2/8jZbOC5XOtu27eH000WAOhkERiVjDt6Iq34+psB92R+aVmOw7AApAg4ZUrOhdPPnSNYZ6Hv6NBy1pbgaqjFY4giK7Jq8eJqqYK8pRdNUhsz6PZV5v1G8+T0kCeKHnENU+ugu+Anaqi3ahc8zCott31E6kglz0DTs1Ztx1JQQEB5/9Ao6SAQoP5CRkYYsS6xdu4biYg8RERYuuGAA27ZdyJtvvobX+0qrqz+lT59+REVFdXk/srOz+ctfHiUr61eiouK5//47uO66a7vsqUhRFH755Te2basDgtG0OoYODefll5/r1mXlLpcLsLQEp/0sFhs1NUXd1q5wYsmyjvRJ0yja+Cs1e+cDEJoYTeygi/jlzXuB34AR+65WkA2vEjuo/WeWtZemqezN/pn8tUvxuuoJSxpE+pSLCU8Zcuyb28lRW0buLyvxOmMAIzr9GlLGZJJ56R+6rI3DcTfZkaRBbcokSQIpDo+zia4+DlYEKD/Rr19f0tL6oKoqun250B555D7mzZtIVdV52O0zMZs3odd/xbvvft/l7W/atImJE2fgcDwGvEpDwzZuv/0uysqquPfeP3dJG5s372DzZpmEhKnIsoyqqmzYkE1w8C4GD+6+E1oDAgIwGJy43c42J/Y2NlYwYED3DJMKPcNgDiB1zJkkj2o+/HJ/PrxB5/yeLfPOQlN/j6YmoTO8R1CURtzgiV3ehx0LP6Fo015U73+BZCpzP6Zm72Oc/vtHsYR0/sFS9XnJWbEcTZmDOah5M6/PU8ruVf9m0MwITAHBx6jh+AWEhaOp24GxLWXN2TN2YQk+vcvbEwHKj0iS1BKcACIiIti6dR0fffQxK1asIz09heuv33hcCxUqKiowmUwEBx/+l/ehh/6B0/lXYP8CjDgcjm944olx3HXX7V0yvLdxYyHR0ZNa3pZkWSYycgAbN67u1gCl1+sZP74vixatxWYbgNkcQG1tCQEBe8nIODWTcJ7sDk7UGjtgHLboZIo2LsPj2Epk30lEpY/ucEJXVfHicTRitNqQdYd+fXqcjRRtWISq5AL78xv+BdVXyZ5ff2Lg9GuP8yc6oLFyLz5XGmbbgUwTemMsPtc46opziU7vvjm1kLg0rKHzcNR+i9EyBk1z43UuIKKvDXNQx3MPHosIUH4uICCAG2+8gRtvvOG47l+7di3XXXc7e/bkomk+JkyYygcfvEZMTEyb67KzN6Bpfzno7jTARlFREWlpHc8gUVNTQ2lpObIsk5AQh9vtIySk7ZJUg8FEY+PBx1N3vf79+xEYaGHjxl00NroZOjSMIUNO7/BxIELvFRAWR8aUK47rXk3TyPvlG/J//QFN0yPJKn3GnUfq2HPbDIHbq4uR9f1QlbbJdzV1OvUliw6utl1Un5f60lxcDfWYg0NQfV7gcJtvg1C8e4+rjfaS9QbSJ02nImcDNXuzkfU64gYnEZE6oVvaEwHqJFZSUsLUqbNoanoRuAxwsWzZE0yZMott29a1+WD17duHwsJsoPWy1HJ8vhqio6M73HZ29iZ+/bUKnS4BTfMhy6sIDPRRVVVIVFRyy3VVVXvp37/r59MOJyEhQZw1JRyX/LXz2PPrNlRvFtAX2EXeqovRm0wkjTiwIs8SHInqy6P5GNVWMzJSFtbwji/+8Tgb2bVkPu6mdGAImpaL0boNVQFVOQ9Z19yGpvnQtLXYotI782O2i95kIW7wOOJOQBYlsQ/qJPbWW+/g9f4OuILm498D8PmeoKjIzfLly9tc++CDf8Jq/RuwBNCAfKzWa5gzZ06Hj1yvqqri11+riYubSFxcOvHxAwkNPYOaGg1Z3kJx8Raqq4spLt6M1ZrH8OHdN7wnCF1hz5qfUL3v0hycANJRvW+ye9WPba4zB4UT3uc0ZP21QCmgAnORdc+QOqbj+/tKNmfhtp+J2XYdZtsELMFz8DqnYbb5cDe9iKtxFe6mLFwNLxPRRyIg4uR6ABNvUCexnTvzcbsPHo+WgNMoKChoUzplyhTee+8l7rzzVioqijEYDNxyyy089dQjHW63sLAMvT6pzao5szkAvT6RiRNDcbt9VFeXERkZRGrqJHFmlODXNE3F6ywHDn5lGIrHXn7I9aedfyPbF3xM6dZ0NE3FbItj0IzbsUWndLBdjZq9JZgCbm9TbgwYj881n/5nDaGmYB2aqhESH48tJvWky44iAtRJbMKEkXz77Q/Y7a0zT7hQlMVkZv71kOsvueRiLr74Iurr6/etfDu+ne6yLKFpyiHlmqZgsVjo21ccqSH0HpIkYw3th6P2Z+CcVn/zIwERhw6p6QwmBp8zm4HTr0LxetCbrMcdOCRJRtO8SK2PgNd8SLJMYHg8gV2878jfiCG+k9jVV19FePguDIbbgc3ASiyW85g+fQqDBg067D37k8geb3ACSE6OR1UL8HgOHIzY1FSL1VrbLfu3BKG79Z96CbJ+NvA+kAe8g6y/hf5TLzriPbLOgMEc0IngJBHRJxF30yI0TQOa36rc9kVE9Dm5hvKORLxBncQCAwPJylrOgw8+wTffXIzZbOXWW6/hT3+6s1vbDQkJ4cwzU1m6dCmaFoOqerFaqznnnBHo9eJXTuh9ItMyGfE7I7kr/oW95h4CIxLpN/FOQhMHdGu7sYNG4qhbSGNlLhJ9QcvFFu0iduBZ3dquvxDfFie5yMhIXn31BV599YUT2m5GRhpJSfFUVFSg0+mIjh7WqbcyQehp4SlDujQbRHvojWbSJ5+LvboYt70OU2B/AsLiTrq5piMRAUroNhaLheTk5GNfKAjCEUmSRGBEAoEn2Qq99hBzUIIgCIJfEgFKEARB8EsiQAmCIAh+SQQoQRAEwS+JACUIgiD4JbGKTziq3NzdZGfvob7eSUJCCKNGZRAZGXnsGwVBaOF12SnZmkVNfjFIMhGp8cQOGoneaDn2zacw8QYlHNGOHTn8+GMRmjaS6OhzqKzsy9dfr6empqanuyYIvYaq+Ni1bD6VeQPRmx9Bb3qA8l0p5K5YgKapPd294+a216F43d3ahniDEg5LVVV+/XU3sbETMJmaz0wKC4ulstLL5s15TJoU1sM9FITeoaF8D866RCzB01vKzLbzsVfn01RZSFBUx/cKVuRms3vVlzRVFRIQGkvaxCuI7DscVfGx6Plr950Z1cwW04dxs59uV70FWT+yZ83XaJpK6pgLSRk9C4CmykKyP/87luAIMi//Gzq9kb3ZP9J3/O863PeOEAFKOCy3243TKRMW1vZAv8DAMMrLc3uoV4LQ+7gba4GMNmXNmSD64rZXEkTHAlR9aR4bvnqGgdNvJDJtJGU7VrH+i6fI/J8H0BvN6E0BnPGHA5ljJLl9A2Wl235h15L3GTzrdoIik9nw9T+xhkYT1W8UxVuWEjfoDGqLdlKzdyvBMX3Qm6zI+u7NDiOG+ITDMplMWCwqbrejTXlTUw1RUYc7zVMQhMMxBYWC1Pahrjn5ax6mgJAO11e8aTERfYaTcNpUTAHBJGfOJDgujYpda6kr2klo4gAM5oCW/7V3nit/7VzihkwmdsDpBEYkkHbGZRRkNZ935XU2YgmJwWKLwOtopGjTEuKHTO5w3ztKBCjhsGRZZsyYPpSVZWO316OqKrW1Zfh8Oxg6tOPHvwvCqcoWnYrFVoCz4WdUxYGqNOJqmEtAeC2BkYkdrs/jaMAS3HahkiTpkGSZuuKdNJTtZvELc1j43LVs/O4FPI6GY9apaRqNFflE9BnWUhYcn059aQ4AepMVr6sRj6sRWW9A8bgwWoM73PeOEgFKOKL+/fsxY0Y8kpRFefkPRETkctFFwwkLE/NPgtBesk5P+uTpRPbdhtf1ID7Xo0Sn7yFtwtlIUse/gm2xfanavR6fxwlAY0UBdSW7iOybib22jOj00Yyb8w9GX/kI9qoiti94+5h1Kl4XmqpgDYluKTOYAlA8LrwuO9EZ49i96kvs1SW4GqqIHXhGh/t9PMQclHBUaWl9SEvr09PdEIRezWAOIDlzEkkjms916kw28uTMmdQX72LVf/6XoKhkqgu2EBSVQnjKEMbP+UfLdZbgSDLOvI7sz55A8brRGY58crUk6wCQ9caWMp2h+f8rXhehCRlMuv11ZJ2enGUfU7N3K9mfPUGfcReROLzjR9m3l3iDEgRBOEEkSer0URk6g4nhl/yF0Vc9SnTGGBSPk4wzrznstaagMDRVwW2vO3qdeiM6gwlPq+u8++afZV3zQgi90UJ1/hYi+gwnb+Xn9J86m9yVn3XqZzkWEaAEQRB6IbMtnLLtq4juP46wpEE46yr45T//i6r4Wq6pK9qBrDNgDgo/Zn0h8enUFu1o+XND2W50BhMGy4FFUTUFmwlPGYLH2UBkWiZeR2PX/lAHEQFKEAShF6ou2ELN3q1knHktAJaQKDRNYdv8t2go20PZ9lXsWvohSZkzkHXNszk+twNV8R62vtiBEyjI+gFnfSWKz8Pu1V8T0Wd4yzxZY0UBgZHNS+IN5kBqC7e3CV7dQcxBCYIg9DKaqrBj4Tv0GX8JFltES/nwi//C1p9eZ+2HD2C2RZAyahYpY85v+ftf/vO/pIw5j+TMmYfUGTdkElV7NrLyzTuRdUb0JjNDzrmt5e+r9mwgeeQ5AKSMOZ8NXz9DH7FRVxAEQWhNknWcfv2zh5QHhMUy+sqHj3jfpNteOXKdksxpF9xFQ9lu3E21hCYORG86sIcqdcwFLf+/z9gL6TP2wuPrfAeIACUIgiC0sMX4z6pdMQclCIIg+CURoARBEAS/JAKUIAiC4JdEgBIEQRD8kghQgiAIgl8SAUoQBEHwSyJACYIgCH5Jaj44q50XS1IlUNB93REEv5esaVrksS87MvE5EoT2fY46FKAEQRAE4UQRQ3yCIAiCXxIBShAEQfBLIkAJgiAIfkkEKEEQBMEviQAlCIIg+CURoARBEAS/JAKUIAiC4JdEgBIEQRD8kghQgiAIgl8SAUoQBEHwSyJACYIgCH5JBChBEATBL4kAJQiCIPglEaAEQRAEvyQClCAIguCXRIASBEEQ/JK+IxdHRERoKSkp3dQVQfB/2dnZVZ09UddmNGlRZmtXdUkQep28xrp2fY46FKBSUlLIyso6/l4JQi8nSVKnj2qPMlt5duSZXdEdQeiVLlzyVbs+R2KITxAEQfBLIkAJgiAIfkkEKEEQBMEviQAlCIIg+CURoARBEAS/JAKUIAiC4Jc6tMxc6D1cLhefffYZa5cvJyktjWtnzyYmJqanuyUIvU65087S0gKaPC4GhUczKjwWnSye7U8EEaBOQrW1tZw5ZgwRJSXMstvZYjYz9IknmLd4MaNGjerp7glCr7G2qpRXtq7lWk1lpKbxQXkhC4JCuHfYGRhkXU9376QnAtRJ6OnHH2d4QQH/8XiQAFwuzgRuufpqsnbsQJKkHu6hIPg/r6ry+rZ1/KAqjNtXdqeqcFZjHYtLC5ge36dH+3cqEO+pJ6G5n3/O7fuD0z6XAwV791JSUtJT3RKEXiW3oZZ4aAlOADrgdlVhfXlhD/Xq1CLeoE5CZpMJ+0FlHsCnaRiNxk7X39DQwPz589E0jWnTphEcHNzpOgXB3xhkGTugQZuHPTt02fBeob2B3MY6osxWBgaHi9GNg4g3qJPQlTfdxGMWC+5WZc/qdIwdOZLIyE7lOeXrr74iJTaWd66/nnevv56U2Fi++PzzznVYEPxQn6AQvHoDH7YqqwOelnWMj0/tVN0+VeWFTat5LGsJpbvW899Nq7h37ULqPK5O1XuyEQHqJHTHXXcRftZZpFmt/N5qZXRQEB8lJPDmxx93qt6ysjKuv/pqFjoczGtoYF5jI0ucTm667joxdCicdGRJ4k9Dx/FnvZGJOj1XyTrSZJn+McmMjYjrVN1z9+Yg15ZToCp8qijsUHyc52jire0iGXdrYojvJGQwGPjku+/YsGED69at4/KkJM466yx0us4NS3zxxRecr2mMaFU2DLhY0/jss8+46667OlW/IPib1MAQXj59JtnV5TR63TweEkmsNbDT9a4o2c37qop5358l4CE0omsrcfq8WPSGTrdxMhAB6iQ2bNgwhg0b1mX12e12QhTlkPIQrxe7/eBZL0E4ORhkHWMjO/fGdDC3qhJyUJkFkJHwqiqWLm2t9xJDfEK7nXvuuXym11PTqqwY+FCvJz05maKiIlRV7anuCUKvMTwilldbLYjQgGeACL2BLTXlVLgcPdY3fyIClNBugwcP5pqbbmJUQAD/kCQeAs4zGrl13DhGu90Ufvcdq+bPx+fz9XRXBcGv/a7PQL41mjlf1vE6MA34CYn7LAGklxfyy671bK4u7+lu9jgxxCd0yFPPP885F13EZ++/T21eHi+efjpnjB6NJEkkA1l5eeT37Utav3493VVB8FshRjP/GH02S8sKeK+6lFiHnUdsoQTom7+SM1WFp4vzSAsOO6Xno0SAEjpEkiQmTZrE8OHDyf7wQyYkJLT5++TgYHLy8kSAEoRjsOj1zEzoS6CsY3hVcUtwAgiRdfTHQ6G9kfTgsB7sZc8SQ3zCcdHpdPgATdPalHt9PnRdsBlYEE4VelmHUzu03KGB/hRPSiveoE4yLpeL3bt2UVtYiMlmI2XAACIiIrq8nYCAAMzJyeSXlJAaHQ00B6ecpibSMjK6vD1BONEK7Q3srCrF7fUQHhTC0PAYrN0w3JYWEs7y8r1kqgqh+zJU5Hg9FOgNjA6wdXl7vYkIUCcRl8vFim++Ia6ujv5BQTSVlbF+yxb6n3suiUlJXd7eiEmT+PXnnykqKsIqSVRJEglnnEF8fHyXtyUIJ9LW2kry9u5ipiQRptOxyV7P97UVzEob2uVBKtJsJSMxjaeK8uiveXAChXojk1MHiDeonu6A0HXyduwgvq6OgfsCRLjNRqjDwZrly4m/8krkLv5lDwgIYMpFF1FVVYXb7WZAWBhWq7VL2xCEE82nqmwq3s2fDQYidc1fkX30BnA52FxTxpioxC5vc3BYNH1tYRQ5GgmVZcZYbad8cAIxB3VSqc7PJ/agxK02qxV9U1O3baSVJInIyEgSEhK6JThpmnbIPJcgdKcaj4twVWkJTvsN1RuoaqjrtnYtegP9bGGkBoZ0S3BSe+HnSASok4jJZsPhdrcpUxQFjyR1SRbzE6myspLrLr2UAJMJs8HA5bNmUVgojjgQup9Vp6dO0/Ad9IVeoyiYDL3rcwTwW3UZ96yZzyVLv+amFXP5tmBnrwlWIkB1gZqaGt555x1effVVCgoKeqwfKYMGsdPhwO5qzoisKApbS0uJGDQIk8nUY/3qKEVROHv8eMK/+Ya9Xi8VisKAn35iypgxOJ3Onu6e0I3yGmv5viiPVRXFeA6TVutECDQYCQkJZ57biXffF3mlovCjppEREdsjfTpeW+uqeGXLr/zD2YQXWODzsjl/B1/s3tbTXWsXvwtQZWVlrF69mqqqqp7uSrvMnTuXtIQEfvjjH1n35z8zon9/nnnyyR7pS3R0NKnTp7PCbmdZcTELy8txDxrEaWPGtFyjaRput9uvh81+/PFHzOXlPOvzEQEEAw8rCumNjXwujvZoF03TyG+qJ7exFsWP/1vvp6gqL25ew3O/LUfO28zaHdn8cfWPFNobeqQ/ExPS2BUSwcNuF8+4XPxTVemblE5iq1V1iqqi+Hlqr+/3bOMfqsL5NB+2OBT4QlWYV5TbYw8AHeE3iyQ8Hg+3zZnDV199RT+TiV1uN7Nnz+af//53p7Nwd5f6+nqu+5//4Senk9H7ykqAUU88wdQZMxgxYsTRbu8WqX37kpSSQmNjI2azGbO5OV+ypmk8/8wzPPfUU5TX19MvPp5H/vlPLr3sshPex2PZuXMn41wuDj66bXxTEzu29Y4nv560p6mOFzetQfW6MUsSDbLMbYNGMzQ0qqe7dkQLS/Nx15STqyr7MnyrvKHAc5vX8NSYs0/4QX5mnZ5pyf1pivPgUHyEGc0t80LVbif/3bGeNbXlaBqMCYviuozhRJr9b4FQsaOxzYnAAMlAAFDrcRFtCeiBXrWf37xBPXjPPZR9/TX5Lhe/1teT63KR/d57PP/MMz3dtSP64YcfOEOnawlOAHHADS4Xn7z/fk91C51OR0hISEtwAnjmySf56JFH+LG2Fo+q8nJhIXfPmcO8efN6rJ9HMnDgQFaYTBz83L88MJBBQ4b0SJ96C4+i8Pf1K3jE7SBfVdip+HjP6+G5Taupcfvv8Ojqkj38tSU4NbsBqHM7KXP2XKb8QIORKLO1JTh5VZVHs5dyek0ZlZpGNRpn1lbwSPZSv3wjSQywseygst2AU4Iwk/lwt/gVvwhQmqbxxhtv8G+nk/0v0OHAcw4Hr73wQk927ai8Xi+mwwyfmDUNr8fTAz06PJ/Px7NPP82HDgdDaD57ZgrwgsPBPx54oId7d6hp06YhJSZym9FIEVAG3KPXszc0lEsuuaSnu+fX1laXMljTuJYDx5SfTfOZXcvK9vZgz45O0TQOniWVAAPg0/xnGG1dVSkJPg9PAjYgCHhE00hXfKyuLO7h3h3q/NRB3C/r+Izmo+p/BS6SdZyXmN5lx9Z3J78Y4vP5fDQ4nSQcVJ4KVNR137LOzpoxYwZ3+nzkAmn7yhqAt61W3urA0JnP52PXtm0Ub9lCTXU1dmMICal96N8/kcTExE4Pb9TX1+PxeDg4v8MYIGf37k7V3R10Oh0/r1zJfXffzdDPP0dRVS46/3yWvPBCm7dC4VD1Hjd9DvOFnqaprPXj48RHRCfyQv52pqpqy1Pzd4BsMJJgDWp3PeVOOxvK91LRUEeTCh5jOOm2QEZGhHTJEFyJs4lxh3lTGq/4yHE2dbr+rpYRHMZdQ8fzRO4mrrU3EGUwMTMpnZkJfXu6a+3iFwHKYDAwauBAvtq6ldZf658CE8eO7aluHVNUVBTPvPAC4/70J67xeglSFN63Wpl15ZVMnDix3fWsXbIE886dRDQ4qM9z4fM2sXybjx07VEaMqOCMM0Z2qp8hISEEWK1sdLs5rVX5UmDwgAGdqru7hIWF8dq77/Lau+/2dFd6lUEhEfwdCTvN8wwAPuATWccsP56DOichjScrSxjpaOBSRWGHLDNXkvnLwNHtfkCrdDlYmruJ81QV2eGjXAnnJxSW2dPJrsnn6j5eUgKDj13RUSQF2PhZp0dTfC1vqBqwUKdnYkDn6u4uQ0IjGTJqak9347j4RYACePqVV7hk5kzyXC7GqipLdDpes1iY70dDfHa7nV2bN1O9Zw+mwECShw7lhptvZuKUKXzy0Uc47XY+vPhixo4d2+4PVVVVFZ6cHIaGh7N8ay0JMUNJknVQVYg+MIVNm3Lp37+qU/n0dDod9z/6KJffcw+vOByMAOYD/2u18snf/37c9Qr+JyUwmOGRcUysLOGefXM6L8o6jEEhZIbH9HT3WuQ21LKjoginx0VYYAinRcXzYOZksqpKWVdXRYjJzPMxSYQY2//GvKmymHM1jURFo1xLpq8hhqs0N8/79mA0X8W8one4LcPWqRGJzPAYvjCZudnp4K9a89ve05JEpdHE6F62BL038JsANXHiRJb8+iv/evppft68mcEjR7Lq3ntJS0s79s0ngNPp5JdvviHVbic9NBRnQwPb5s7FMXky/QcP5sGHHz6uehsaGoiQZRobG5GkMOR948Jhso4KRwOyPo6qqupOJ3y99Q9/IDAoiLsffZTdxcUMHziQD595hsmTJ3eqXsH/3DRgJMvCCnmuZA+KqjIiJomz41LQneCVcEeyubqcosIcLtbriZZ1bKur5IeGGs7uN5QxkXGMOc7j1evtjaTp9VQ6PBjk5gPVQyQTwWojii6UarcVu89LYCc22+okiQdHTOaTvM2MrChGBcZHxvFQ3yEiNVE38JsABc0ntr7Rg6vfjmZPTg4JTU30i2v+8ASYzYyxWFiyZg19MzIwGI4vgWRAQAAlqkq00YiqHRjDrlcUzNYgXK5qTKaQLvkZrrnuOq657rouqUvwX7IkMSUmiSkxXZ8guLMUVWVrWQF3G01E7ts+coZOh+Jysr6yhCmdmBsJNFspaqwhUAan4kUnmV4n2k0AACAASURBVLFrXhrRE4gOWXJhlDufHTzQYOSG/pnc0D+z03UJRydCfjvVl5YSGdB2z4DJYCBQUWhsbGx3PQdvRI6KisIXH0+F243ZXE99QzWFtWVUBkeg0xmwWitO+ezgiqJ0eGNxQ0MDOTk57N69G5fLfxcHnGoafR6sirclOO2XbjBQb69vdz1tNiLv2yw7KCqeb1WNJr2KVymkTrXzjVKNYjmDOs9KhoeB0U/3VJ4oHd20ragqOQ01rKsqJa8HNn371RtUe23cuJHnHn+c7Zs2MfC007j7b39j6NCh3dqmJSSEhr17iWyVjFVRFOyAxWI55v0ej4fb58zhy6++op/RyC6Ph2uvvZbnXnmFsdOnszU7mxq3m+3bNuIKSCI6LhGzeStTp2b2ujx6XeW3337jL7feytKsLIJMJn7/+9/z+DPPHPPfe+uGDZSsXk2cpuGTJJYaDJx27rnExoo5gtZUTWNxWQGrinfjVVWGRSVwbmIaZl33fS1YdQaaJJkmVSWw1ZBYmeLDam3f201+Uz0vbl6N6nFjkiTqZZlbB45iWFg0w1MH8m1JPiW+Wiq8jXiM6QRIvzI41M3ZsdHd9WP5NU3T+L4wh+8LdlHh89DXHMBlaUMYfYyh1Cavh592byXe5aCPppEjSWy0BjIzdRAW/YkJHb0uQK1YsYJLZszgHqeTWzWNlbm5TJ03j28XLGD8+PHd1m5q//78umkTtvp6IoOD8Xi9bCktJWLYsHYFqIfvu4+SfRuRbS4XNcDFH3zAP5OSuOf++8k8/XSGjxuHqqo0NTWhaRohISEnfAe9v8jPz2fGpEn8vamJeUCZ08mf//MfZufn8+n33x/xvqqqKspXrWJybCyGfR+iVIeDVT//TOTVV6M/QR+s3uCN7VlUVZbwoKoQCLxUsINHK4p4ZOSZGLppPsWo05EcHsenFYVcajIRJMnsVXx8p2pktmPuyaMoPLl+Bc/4PFxD816pxQpcsnkNz46dRh9bKKlBIfg0Fa+qUutxEaQ3YjOGd8vP0xt8nb+DTXt38bOqMASY77Jz3bZ1GIeMZVjYkYP2r2UFnOFyMN3c/P12tqbxnb2RrIpCJsSlnpC+97ohvvvvuIMXHQ7+rGmMBf5XVXne4eD+O+7o1naDg4MZdsEFbLFa+amoiEU1NehGj2ZYO5bBa5rG66+9xkutNiKHAc87HLz+r3+1XCfLMnq9npCQEEJDQ0/Z4ATwygsvMNvt5nrARHN6lg9cLpYuXkxeXt4R7yvdu5dkvb4lOEHzkSOhTieVlZXd3u/eotDeQHZlCUtVhQuAqcDXqkqw086abt5wOjYmEW9MEk/4fPzN7eJNWc+glAFt8twdSVZ1GQM1tc1G5DOBSzWNpaXNiZolScIg67DqDcRbg7AZe0+i5K7mU1XmFubwmapwGs1f+DOAF1SFuUdJGKtpGsW1lUxqlWRakiQmm0wU1lR0e7/361WPk5qm8cvGjSw4qPxi4PcbNnR7+1FRUURdfDFutxu9Xt/uHIGKolDncHDwlHUK/r0RuSft3LCB2V5vmzIzMMxoZNeuXfTte5TJ9MMF9lM42B/OjvpqpnFgrxQ0f+Ffpiosqq1gQnTXH8q3n06WGReTxKioBDyqgkWnb/fDWL3XTcphNiL31VRW+/FG5J7S5PMgaSr9DiofBxS3Y2OxpkHrpJgnOu1wr3qDkiSJmJAQcg8qzwViQrpmpVt7mEymDiWw1ev1jBsyhIPzcH8KTPLjjcjdqbq6mm+//Zbly5ejHiYj9KBRo1h+0NybA/jN7WbAUTYXxyUnU+D14vX5Wsrq7XZqzWYiIyO7rP+9XajRTM5hgsIOSSbYdGKSnuplGave0KGRgkEhEfyAROuvVgX4RKdjUJj/bkTuTrsb61hTWUz5YXIWBumNSJLMwe9Ky+Gob6ySJJEQGsUyz4Hz5TRNY6nHTeJRhgW7Wq8KUAC33XEHt1ut7B+sqQD+YLVy25139mS3jumpl1/mDquVJ3U6FgMP6XQ8GBjI4360EflEefapp0hLSOD1a6/ljlmzGJCczI4dO9pcc9udd/Kx2cyLkkQtsA34ncXCObNmkZKScsS6w8PDiTn9dJaWl7O1uJgNxcWsttsZNmOGmH9qZVhYNMU6PS/R/AWvAT8Bn0kSU2KTe7ZzR5EUYCMzMp4zZB2f0pwOabqsQxcQTGb4qbUIptHr4aGsxTz/2zI2bs/m3rULeG3bujYr7XSyzIUp/fmdrOMXoAn4Arhb1nFB6sCj1j86JolVlgBecTn5yeXkJbeLDVYbo6IPTkrXfaSOLN8dOXKklpWV1Y3dOTZFUfjLHXfw9ttvk2g0UujxcMMNN/DUCy/47bEc+23dupV/Pf00OzdvZnBmJnfecw/9+h388n1yW7p0KbPPPZcVDgeJNH8xviVJPJ+YyNb8/DZP01u3buX+O+9kwYoVhAYEcP0tt3D/ww+3a1VjQ0MDFRUV6HQ64uLiuuzARkmSsjVN61TuqTRbqPbsyDO7pD+dUeJo4t+bV1PlcmAGVL2emwaO8utjOaB59eGK8kJWl+zB12ojcm9IftqVXti0mgE1ZbykaehoDj4zZR0ZqQM4Lym95TpN01hYks+8gh2UeVz0swZxSdqQoy6Q2E9RVfbY66nzuAk3mUkOCEbuguHyC5d81a7PUa8LUPvV1dWRn59PSkoKIe0c3mtoaCA/JwdXfT2h8fEkp6aesku4e8rsyy4j8/PP+WOrMg0YFBjIOwsXMqbV4Yr+6GQKUPuVOpvwqioJ1qB2ffmomkZOQy2FdZXIkkxqWCQpAcGn9KKeE83p83H9yu8p0VRaZwBcAVxvtvLMuBk91bV2aW+A6rVjHiEhIQwbNqzd15eVlbFp7lxSNY0os5mynTtZERnJGeedd8KPQ1cUBVmWT8kPdENNDQfPBElApCzT0NAzp6ee6mItge2+VtM0FhXmYK6p4EydDp+msaS2nLLoJMad4MwVmqahgt+kcDqRvKqCjubjPlqLAuyK7zB39E69bg7qeGiaxtZly8i0WukXG0tMaCjDEhKIrqwk96C5j+70888/k5ExEr3eQGhoHI888iSKHx5y1p2mXXIJ71ittF4WsQvY7PUy9hRdMNKbFDoa0WorucVsYbjRxCiTmTtMZkrKC6l1n5hVdF5V5Z3cHVyxYj6/W/oNd61dy/a66hPS9uF0NMtJVwgyGIk1W/j6oPK3kBjmR0mBO6tH3qBKSkpYunQpNpuNadOmdeswm6Zp/Pe//+WzBx8kxW5n0LBh/O6KK4iOjiYhJISNeXlw2mnHrqiTVq1axUUXXYvT+SZwLvX1O/nHP26moaGJZ599stvb36++vp6srO3k5VVhsRgYNiyJQYMykI9jY6bb7WbLlp1s316GLEsMHhzPwIHpR12MMHvOHD566y2m7dzJ1XY7pbLMS2Yzzzz/PEFB7T/3R2j+3d5RX0OZy05qYHCnj5I4lkavh09yN3NGbSW/6WRirDYSA2yYJZkhQJGjkdATcErrS9u3saYqEY86F0gk3/4FD2+8mX+OHNOuvVRdQdM0NtZWs7zcTZ1HJd4qc1ZsIMnH+d+g0N7A0rImSpwakSaJSTEW+gaFHvF6SZKY3T+TGzf+wmpNZbim8YMss0Rn4LFjLH7oTU74HNRTjz3G0088wVSjkUpJIk+v59v588nM7J7Ei08/9hjv/P3vXOd0cimwWpZZZLHwz5dewqfTsTsqivEzum681uPxUFxcjM/nIyIigtDQ5l+y6dMvYf78acDNra4uwWIZTGVlIQEH5fk7FkVRWLx4MXv37mX06NEMacdR6A6Hg88/X4HP15/w8AS8XhcVFdvIzNQzduzwDrWvqipz5y6npCSSyMhUVFWlsjKH9HQnZ5999Iwebrebjz76iAVffUVoVBRzbr2VkSM7d+bVieIvc1ANHjdPbViBx2VnBLBcg34hEdwxZGy3LBZwKT7+unYh/VwOLgJmAVlIGMxWBoVF8qbLSWjqAPrZwrqkPU3TKHc5KHI0YpJ1pAWFYtHrqXW7uGnNcrxqMXAgGMk8wqToj7hz4OAOt9XgcfNbTTmyJJEZHkOA/tiJn7OqKvm+KIogwwWYdRE0+QrwKJ/x+3564jtwwCI0B6d3cxUM8iVY9Qk4lXKcytdckeIiI/jo/55lTjsLi/KodjaSFBzBWXGpBHUiW/uJ4pdzUMuWLeP1p55im9tNrLt5ff2XwCUzZ5JbUtLly4AbGxt5+u9/Z4PTSTngBq5WVTwuF19+/jn9zzmHvmd23WR1ZWUlc+dm43ZHI0lmVPU3Ro2KYOTIoWzfvhN46KA74tDpwigpKenQar7CwkJmTpyIubqawarKg6rK5OnT+e/nnx/13zA3Nx+XK5G4uOZlxDpdAPHxI9iwYSFDhzqwWtu//6WkpITiYjOJiYNayhITh5GTs4zhw49+fpXJZGLOnDnMmTOn3e0Jbb2z8zemOhp5UdOQAA9wQV0lX+fv5LI+Xf8EvaxsLxkeNx8Cj+9r71w0PnLZyXYGsNtg4pIueoPTNI2fS8pYWxWMqp2JLNVikrO5sk8gdp8Xg5SKl7ZvSipnkG9/p8NtLSrJ5785G5gsSXiReEvTuH3gSMZEHjlBs6KqLC33Emy8DJOueYFWkKEPddoFrCj/hP9J7ViAWlJmxyBfhc3Y/B0QJKegky5jYenrpNu0o85Vx1gCuLpf9+Yh7UkndA7q/Tfe4E6Hg9a7FS4BIlwuVq5c2eXtbd++nVSDgSTgNJqPY18E6BSFTzduJHryZBISEvB4PHz88cf835/+xCsvv0zdcWR3UFWV+fPXYzKNIiFhOPHxA4iLm8zatXWUlZVx2mmDkaTlB91VgKrWdjhb+Y1XXMHlhYVkNTbyrt1OntNJ2fz5/LtV2qTDqapqwmJp+0Sm0+mQpGDs9kM3+R1NbW0DOl3bICRJErIcIRY7dDO3orCmuozH9gUnACPwhKqyojS/W9rMq6viIlUhDJgNvAI8B3wNvKwqTOkzEIOso9Ll4Mv8Hbyfu4mNNRXHNT+T11jHr5XxhJn+QJRlAhHm85HlG/kiv4EosxWvthuobXOPzCL6BnVseLHU2cT7ORtYq6p8qyj8oPhYrCq8vC2LhlYbVA/mUhWcPnNLcNrPqouj3NmhLgBQ7FAI0LfN3GHRxVHtlk549nB/c0LfoByNjRzuGStEkjr8BdkecXFxFHg8OAELzek9GoEsICozk/6DB1NTU8PUsWMJKS1lelMTy61WnnjgARasXMnAge1/Eq2pqaGx0Up8/IGklDqdDoslld27S3n44f9j8eKZOBwhwPnAdqzW27n77rs69OZSWVnJmqwsvm21uMIMPOBw8H+vv85dd999xHsjI4PIyakhNPTAJGrzIo16AgPbv5ILwGYLQFUPzW2naXUEBIiMDd1J3Zfq5+DfmhCavzy7Q5g5gE2SDJrKUOApmjO4vCnruHfsAEbHBrB0bwkPr13HZZpGmqrycekeVsdE8uSZ49DJ7V9pt3yLm+DA8QRbDgy1BZFAhT2B8OhGzuubyA+7z8WlvEJzwrBPMev/zR3jxpHcgSmoRRuLuBqNjFZlmcAMGXK1Ei5JPnxCVEXVE1HkRC/VYtYfmCeqdpYwIBxSO7jPOa1MpsJRRpAppaXM4a3Cpmqn5ArF1k5ogDrnsst4bfFirrHbWxreAWR7vUyYMKHL20tISGDKlCnctngxL7jdBANbgH9ZrXxx770APPHQQ4wsKOANj6f5adTh4FWnkz9cdx2L163rkn74fD5KSsq59dZr+fHHf5OTczMREfHcc88d/OEPt7J79x62by9GVTXS02NIS+tzxE3HbrcboyRx8Ch5EM2n/h5NWloK69evoLzcSkREAh5P8xzUyJFR7crI3lp8fDxhYTmUleURGZmCpqmUl+eSkKAQFeXfGz17msnY8S+xtgwM2mrj/Zp6ft+q9FVJYnJybCfrPrzZ4alcWZLLWT64APABH0tgDLAwMiYSc2oyj3/6Az8pCqP33XOPT2FCVS0bY6O4dHL797dF6Fdh3RKDLbTtD9JUFYX19HSmJMbiy9rBL1tm0OCoZ1TGAJ655W7iIkL5fvV2Nu9uJDbcxKxxfRiQfOTRCavLg7xh+yHlNkkmaOgA+l5w1hHvvanvVl7+ZikW2+8IMEdT25SL2buaG288n74JHVtFd9OwPTz63i9IQX0ItCTgcJXjqlvG5WkmJJ8IUCfM5Zdfzidvv83p69ZxdVMTFTodb5pMPP/SS9hs3bP65u1PP+X2OXNI+v57bDodOquVF/7975ajOb7+9FPm7g9O+1yvafxl40bq6uravQk4LCyMoCAHDQ3V2GzNb1GK4qOw8Ffuu+8vOBwJ+HwDkKQSZs48ny+/fB+9Xs/KlVls2OAlJCQDSZJZsGA3BQVrOPvs8S1jzz6fD0mS0Ol0xMfHE5+QwJe5uVy6r20NeNlo5LzLLjtqHy0WCxdeOJZ167by9dcfkpX1G+HhOtLTb0DTjjzWrWkaLpcLs9ncco1er2fWrHGsW7eVnTt3IEkwZEgMI0eOPSX3d3WEzmwiJP34T44FeOTKQGa/+QlrFIWRPoX5Bj3ZZhOfXXouIbaOvQ23Rwjwemgo934+j5ubHLg1jcyEGL59vDkwLFm/lTSJluAEzVnob3O5+T47l8vveazdbU0yJ7H40WWo4TPR6Zr3KNbV76TKt5dz7/sYvf4cZCkeRc3h/bdf4/xZ51BRWcn//fVVGhrPItg2hJLKMlZ9/h33/Hk448c290rTNJxOJ2azGVmWOUcfxzXfLuIep7NlZKcc+EanY+mNdyKnHvlIiRmjz8Vy2ire+/AbVq3fis/nYvKEdALGnoMcf+RjQ3w+H6qqtlm5PGoM3Jexjv9+8CPFZY1ERFq469apTDQMZ+93C9lT0O5/upPOCQ1QBoOBb+bP57vvvuPnb77BFhrKghtuaNcKtOMVFBTEe198QV1dHbW1tSQlJbV5OzHo9XgOusdH85d+R1InybLMtGnDmTdvHUVFUTR/PEv5+ONnqKz8I6q6f+jNzcKFM3nllVe59tpr2LSpnsTEKS3LvIOCwsjN/YWhQ8sJDAzk11+3kJNTgyzDgAFRjBo1mFfff5/zzz6bRT4fg10u5gYEUB4by5L772/Xv8dbb73J3LnZ2O03A16WLfsDN954Mc8//9Qh17/55tv87W+PU11dTEhINA8//Fduv/0WJEkiICCAyZNHM3Gi2vJvIByb3momfPiRE962x6ThA1gzbjjv/biM1UWljOvfl5fOOh1bQMfehDti5vABzLjoLIoqazAbDUSG2DDGNOdlMw4Jw6t7A2i7F8oN2CUj2bXtf2jREocybOpOVix4HE0bjiTVoypr2LR1DV7Pejye/YEjm+tuOouvfhnPkh+WUlgzhaiYmbgA2ZAE9jj+8fq/uS9tJJvWrWXeF4uoq3ETaNMz/YKJjJ08mXEXXc6wrz/jRpcTryTxhsnExTf+gdqQPmTXHq2XEo0hifyw4GtczrPwuKdSVLKWT786m3999DEZg9tuXWmsr+OfDzzI8vnfoipe+g8dzz1PPkafjObfA2P6aG54ZBRejweD0QiaijHYR7IkwbcLTtkg1WtTHXWVh+67jy3PP89nLhf7w9GjOh3rJkxg7pIlHa7P4/FQWlqK1+vF6/UyYsQEXK5SaDMot4CBAx9k7twP+flnOwkJbQN0SUkOY8c62b69nKamNCIjk9E0lbKyXGJjyzj//EmUlZXx7n/+Q2FeHqMnTeLyyy9v1zDdypUrmTFjDnb7Rg7MYtRgsQxk/fplZGQcGJF/770PuPXWR3E43gfGAL9htV7LM8/8kdtuu/kwtZ/8umKZeebAdG3NRy93VZd6jC/5QCaXNWVurhw/lLfrapm5r6weyDRZ6PPHJ0nOnNjh+hsrS6kr3o3eZGHvb7+w5YdwNPWfba7Rmy5m/OyhlOdU4nPfhMHc9u3FXv0QA6YPZOuPuZiDZmMwx+Nzl+Osf5/B56aQMHQ0ZTvWU7TqZySdnuQJ5xDZdxDtsfC5hyjIGommtV6d+w4Rqa9x4ZOvtpRomsY3991MbVEmqu8Jmgfk38JofYRLn/sIy2GWkl83OglNVcgM9uH9ZS4FJ1mQ8stl5v7o3gce4MLlyxm0YQNn+XxsMBqpCglhwXvvHVd9RqOR5OTmsfM9e/YgSQYO/We24na7MZvNNOdjb0tVm2hoaKK2NoyEhP1PizLx8f0pLKyioqKC2NhY/vq3v3W4f99//yN2+5W0nWIPQ9Mu5KeffmoToB588GkcjtdpDk4AI3A43uHRR//nlA1QXUEzWdt8ufdm+9+MDCYzI+/6J5c/9UdGKQoJqsq3gKXvOIoaQileuuU4W2gefCsrrEJTDz2jSlEs5GwrwGsHV00e+lYPaZrqRHGVs/a7ahT3bOxeK82r/4yo3nNZ/eW/iK+xIklG6HseADmFGjmF7etrwfqVaNprB5VeRVX+zaxc+BuyvnkYz1GWS01RLZrvVQ4snL4dr3sd8//zJpEjzj18AxowKp7M088jGU7JN6lTPkBZLBZ+WrGClStXsn79eqanpjJz5swu2ZOVkpJCTEwke/Z8CfxuX6mG2fwyV1xxATExMYSHb6OiIp/IyGQkSaK6uoSgoEoCAqKQ5UPnEmQ5GIfDcdx9Cg4OwmgsxHPQuKZOV3XIPGBxcS5tZxUAMikv34OqqmJI7zhJsg7JemIyHnSnrOJGAP67di8Au0v1xF3xHDvyf6NE52Nw6lACwzu2heJIgoaOZd2211G9f4aWGaNCJG0e/YY8hdfZxK6lP2CU49EZItFUF67G74jvn0jVniL0gUnI8oHPtGaMx9WgkBBiRj7Ojc25RiseXxXQ+mesR5L1JIYHtdRbXFSLLI1COWhXj6aMR9f0BYmhh67iXbV0C+MnD0aSdWTXQ+bp59HXbIJPvz+lgtQpH6Cgef/OhAkTunwloSRJfPTRG5x99vn4fD/x/+ydZ3wV1daHnzn9pPdeSIAAoUrvIF1QFCxwUQTpYud6VexeEcu18Iq9XuWKKAiKIiBNQEITSEKAQAghvffk9Jl5PxwIaUACJyRint/PD+zM7L3neM6svdde679Mpk64uKwnPNzEE098bD+oHd+f3btjSU09BQgEBmoYNqwvlZWViGI60K6qP1mWkaQC3N2v/Ed/993TePnlnsBc4PwqfifwO5MmfVbj2nbtupKYuAO7bgBV14aFdWo1Tq0AIJwLH3fR2F8lbfy9wX+0w8fxCO5AcJceZCV0Q7TOAsGIQvkF7YZMRu/mg97Nhzb9DGTGvY2p3AWBcvzaBxLcrT/G0goqi06jcbrgSreZz6J3d7li4wQQ0n0IZw8+iWT7EXuyh4igfIKADkNq9OvqF4YsrwWsVHf1K1S/4R506dpK9s9XiaBSXPNqti2BVgPVxPTv35/Tp4/y1VcrSElJZ9iw+5k8eXJVFI+Liws33TQYo9GIJElVkkdubm6Ehp4mPf0ovr6RSJJIQUESHTvq8PK6cjmZsLAwvvrqI2bOHIFK1ROwIssnWLt2VZ2IxTfeeI4pU+ZjNL4PDAX24uS0kNdee6PGdefn7uTk1KQRfJWVlaxatYrEo0fp3KMHd911V6NyyFr5a9NpzHQCO/cjN/FPFEoFgZ2fxNXvQii6T5tovMI6YDWUodToUWnsibtBXbpwcsdqLJUSKl0kNnM6onUNbfo2XhapOm0H30p5wccUpoQiKPojy0dw8/On05iHa1zn5h+BR3AYJRl3IomvAB4gfIhKs4fgrhcCkyRJtM9d3fR6hpmGcv7ISccqS/TxCbqspFJz8bcPkmjJWCwWjh07RVxcKhqNmu7dw+jYsb1DCjNWVFSwY8cOVCoVN95447nzsLps2LCBxYuXkpR0jIiIDrzyylNMmjQJsGv77d59hDNnKhAEJb6+CoYN63ZJmaMrJS0tjeF9+9KlooJBlZXsdnHhpJsbvx840GgljqvBIUESvXrJe2JiHDWlZuPPzHIEhcB/96XiolHx22+x9bqrWgKVhZlkHUugsrAEvYcbQZ2jaxi3q+27PD8NJ88A3PwvktxrNXN69zoy4/9AEk34RPaiw4g70Lvbk9qL00+Sdjgem8UVqEDh58b4BTOZM7QDsiTTy1NG2v8LyQ5w8W3JPMO3p+OZLsm4IvOVQknPgDBmRvW4ZikirUESjUCWZdavX8+HH/6Pykoj06bdzKxZ913zOlG1OXToEAsW/IuEhANoNHqmT5/BsmWvOWTX4OLiwi233HLZ6yZMmMCECXUPcWVZ5rff9lNQEEZwcCSCIFBSksfPPx9i6tShjU78vRz/WriQGQUFvHBOQePJigqeMRp56uGHWfHDDw4dq5Urx2IsJ+3QFgrOJKFzc6NNnxF4BHe4/I1NiCxLZJ84SNqfG7FZy9G5BOMe6OUwA+XsHYzzZc7alGotHUZMpcOIqXX+VlGYyZm9p1DpHkbnGogsmSlJ/4njv22EoY797EosJlacjuewJHE+E+9xSaR7Thon/EOJ9nD84vJqaD1IABYtWszddz/N5s0T+OOP+3j88bUMHz4Bq9XabHNKSkpi9OiJxMfPRZLewWTqy2efbWTIkNHNUn+mNgUFBWRnq/D3b1u16vLw8MNsDiY1Nd2hY8myzI+bNvFIrdpZj4oi6375xaFjtXLl2IxlxHz+HCl7VZRmvUBu4i0c/HY5mQm1NSivLUm71pKy7xQ2y0qQF2Aq9yfup8/IPNr4NJKmIP/0KQTFBFQau0qpoNCicrmZ7GNZlJeVOnSsQ4W5jBIEqqeJuwOzJZEDeRkOHcsR/O0NVEpKCh999BmVlbuxy2DejsGwkYQEAz/++GOzzevtt9/HZJoLrAB+AhYgy89w+HA29923oNnmdR6TyYQg1C0RolK5UFHh+MJ1KqWyTkK1GVA7wN3ZimMonr0iHwAAIABJREFUjN2CxTgWSfwae2DNo0i2zST+9g2SrXkWe5LNSuqfm5BszwLTsZfoeBHk+0n49SvKcs40y7yqY64woVB712gTBA0IHhgqKhw6lkoQqKmbY8cEKFtg4FPLm9E1ZteuXahUY4Hqh4RKKiqmsmHD9uaaFgkJp88JuRqAjdh132cCh/j229WkpKRctg+TycTx4yfZseMg8fHHryo8vTaenp7IckGdisAWSzYBAY49cBUEgam3386/1eqqSCYZ+LdGw9QpUxw6VitXTnn6aWTx3lqt3ZHxoaKweVbnFmMZyBpgOXaJ2yXAOGApyO9wYsv3DeqnsiiLtMO7Sdm3g+KMRCQHivK6BXgimo/VaJNshSg1RXj5OFZ4uZd3ALtkmUPV2jKBzxRKBvmHOXQsR9AiDFRCQgKP3n8//7j5Zt5/7z0qHLxquBTe3t4oFHVdUmp1Ov7+3vXccW3o06czgrAGcAE+AorO/cUTlWosO3fuvOT9lZWVrF27m507LaSlhbB3r8D33+++olIi9eHi4kLv3v6kp++luDiXsrJC0tIOEx5uJSjo4lpkV8oby5ezv107eru68pBWyw0uLsRHRfHqO+84fKy/Kjabje+//5777ryThTNnsnvXtXWtqXROQO3fkgVZyketc7w+YEPQOLljr161DTgJbAKkc3+dRklW3GX7yDsdx4mtseQn96E4cyTJe8o4E7PFYUbKt20X1E4xmEp/xWZOx1IZh2j6hE4j+9tljxyIi1rDA9G9GaFQMlmpZIZCSReFggnhHYl0bZju6LWk2YMk1q1dy/x77mGhxUIvUeSHHTv4ZNkydh06hLt705awBhgzZgxa7QPAV8C9gAAcQKX6L3Pn7m3y8evDYDCwffsfyLIaGAzEAP8GNgC9UakyLxspFxubSEVFBCEh5/OoAikocGX//uOMHXvpirfnyc/P5733PmLnzj9p3z6MRx9dQOfOF2Rgevfuhr9/OsePJ2O1ivTq5U+7dt3q5EhVVlby7jvvsH7lSrQ6HdPmz2f2nDmNikb09vZmf0ICW7ZsITExkYnR0YwcObI1H+scoigy7dZbyd+3j9mVlZQJAvPWr2fWP//Jvxqg0egIvLsNwlzwAqJ1KBAG2BAUz+AWEFkVrXatKc1KQpYl7InyTsBT2HdT64A0VJpLv5RtZiMZsafQOi1GobInV8tyN0oyP6Is5wweQQ0rNFqccZK0QzuxVFbg274jId1HVIXBq3UudBo1jryko5RlH0LjpMUloC3BXfvU6edkUQkfJySSWVFKoLMbE9p0pN0lSsPXRz/fYKIH+rK/IAuLJPKGdyB+upYZfdmsBspqtfLgnDmsNxrpf67tHoOB6RkZLF+2jGdfqF2B1vFoNBq2b9/AhAl3UVT0KgqFK5DGl19+Qrt27S57f1PwzjvvcuqUH7Ab+ya3AlgNzAYeQ6M5y5gxYy7ZR1JSPj4+NfM8vLyCSEmJa5AKREZGBj17Dqa8fBQm0wz++COBlStvZO3aFYwdOxawu97CwsIIC7u4a8BisTBm0CACT55kqcmEAXh10SL2bt/Ol999d7mPogYKhYKxY8dWjd/KBTZs2EDWvn3sray0p4LKMtMMBqL/8x+mzZx5TULxXdv0JHJgMcl7uqBQdkIS03D1DaDHpAebfOz6kGWJ2B8/Qha/BSZgP7Wcjv13tByFahOhPS9dUbuyOAtZjqoyTgCCoECh7ENp9p4GGai0I1s5ue0nJNu/gFBKsr4m/ci/GTDzOVQae7SrWudCcNcBBJ/LJU4vruuO33f8NAs27OQpUWQYsMdYyZKiPB7rNpCuno1bALiqNYwKbNOoe5qDZjVQx44dw91mqzJOYN+/3Gc288KaNdfEQAF06dKFs2ePcfjwYYxGI3369GnWEPMVK9ZhNL6O3TAdwZ6BHgIUEhDwEr/9tqGGXH996PVqrFYzavWF57DZLGg0igblOjz//FKKi6dgs70OgChOxmDox5w5j5KWdrzB+RLr1q1DkZzMapOp6mh2hMFAu59/JiEhgS5dri5ZshU72375hXvOG6dzBACjVCp2/v470+6++5rMI3LALYTeMILy3LNoXDwcJnV0JZTnpSFatMBNwGbs7j0vIBB4Fb92N9Bu8KRL9qFUaUHOrNMuy6WotZd3v9ksRk5u+xbJdgDOlUaUbHdgKruVjCPbaNPv5kt3UI0lH3/LMlFk+rl/9wVCJJF/J8XRte/Fa1f9lWlW/4ibmxvFNhu2Wu35gHsD6zA5CkEQ6NWrF4MHD272/Ce768sM7AfaAqOAESiV/Xj88ZdrCLpejB49wsjPP14VxCBJEjk5x+nRI6RBxmXjxq3YbLUPvMdQWFhEZmbdH+zFiNmxg0kVFTXihvTAOEEg5jpIVm0puHp6kl+Py7RAEHBtolprF0Otc8YrvHOzGicAQaHAXjznT2Av8CzwHPA8Ss1EvMKjUSgvvUZ39g5C65KPufJwVXqHaM0HduEZdnkPS1n2GRTKDlCjbq+AZLuP3KS6xRIvxb7kVGqb00nAscoypBaQetIUNKuBioyMpH3HjryhVFZFZxUArzg7c98jjzTn1JqV2bPvQqdbAjhj3zkBbCEoqA16fXSDDET79m3p00dHbu5WsrL2k5W1la5dRbp3b1gZew8PTyCrVms5omjE1dW1wc8SGBbGqXoM/imlksDAwAb308qlmTZjBp9pNCRWa/sJSFQqL+sOvl5x8QlFrVcCHwKTAW/ABooTuPpOJT85C9FWO3mhJoKgoN2Q4Wid12EufxNT+fuI1reI7N8Jvdvlk1pVOmdkKYcLgRnnyUKjb9y5T4CLM0m12pIAb5UaxXVaJLTZT5j/9+OPrI6MJNrFhZvd3Giv1XLr/fdz++23X3GfaWlpLFr0FMOGTeTRR5/g7NmzjpvwJTh27Bh33TWTiIjujB49iV1XGEX14IML6dNHj1r9H1SqT9Drn8fV9XueeupRBMEZk8l82T4UCgX9+vVg+vQh3HZbG+69dxBDh/ZpcGDCY4/NwcnpGS5ED9rQaJ5izJjxjQpeuXfmTH5QqfgFe2i4CLwnCGQ5OzNu3LgG99PKpYmOjuaVZcsYoNMxys2Nvq6uPOjlxXc//3zFHgFZltm8eTN33jGD8TdN4av//hdLbRn8JkCWZTKP7iTmyyXs+uhpErd+g8XQ+IRVQRC4YfJCBMVOEOJB+AMUq9C5qnDxaYss6ZAuY6AAdK7eRI+9jU6jO9NheAjdJt6GZ2jDFB5c/cLRuepBeAuqluFnUaheI6z3sEY9z7zJY1moVJJz7t95wHyFkrHBV1eduSXT7FF8YWFhHD55kv3795OTk8On/fpd1co6Pj6ewYNHYzJNx2qdxd69MXz+eV927drMDTfc4MCZ1yQ2NpbBg0djND6OJD3C2bOHiYmZyooV7zN58qX93LXRaDT8+OMqli79gaIib3x8OtGvXz/UajUZGbvw8WnYLgjAycnpiqSR5s6dQ1xcIp9/3hadrjdW6wl69Ijm668bF9gQFBTED7/+ytxp03iwpASTJBHRti2b1q5FrVZfvoNWGsw9M2YwcdIkdu/ejZNez+AhQ67qM372mZf46KO1GAyLAFf27/+UFf9by6ZNax1SjuZiJG79lsz404jWJYAfaYc/J+fkSwya/TJqXd3k8EvhFhBB1I3jyIzPQqkORuvSBY2TGzZzKhpnEZW2Yb8NQRBw8gxo9LMIgkDPux7m0HfLMFd8iKAIQhLjaT/kDrzDG3f++uAd4zizP5YOiWcIEBTkyBIjA8OZHHF11ZlbMtedWOywYRPYtWsCsLBa6ycMGPADMTGbm2zcMWMms2XLCKB6xNIOgoPvJz39xBWJMO7Zc4gjR8y4uUUiCApKS8/QpYvAsGF9m0TUcf/+/Sxe/ApHjhwiNDSCF19cxMCBA4mLiyMsLIxOna78hyDLMidPnkSr1RIRUb+g5l8BR4jF9ryhh7xn+xZHTalJSM/IoHufUZjMJ4Hzriwbzs4D+fSD+5k0ZVodsVjgqgVjzRXF7PrwcSQxherJ8wrVnbQb4kREv8vrR9bps7KUxK2/YTUNRa3rhGjNQZY20m5IF9wDHb/7kCWR1D83k3poFzZzOT4RXWk/bDJWUzlWYwXuge0ua2jTiw0MHN6Fmf3D64jFHku2kmcy4Ktzwln111zk/W3FYmNitgGrarVOZ//+hciy3GRqvQcO7AP+r1brcPLzsygpKcHTs3G5CgADB/YkJCSNxMQUZFlm0KBA2rQJb9QzGAwG/ve/b9i58wAREcHMnXtfVcXf6uzfv58RI27GYFgKvEdJSSzTpz/KsmXPMnfurEbPvTaCINCxY8er7ud6QDAbUKXGNvc0Lsnu9b+hUo7kgnECUFFZeTeb1vzIbbfcDAjIkszM/uE1jNTVUJqdjELZF0msqUYi2W6n8Oz7RPS7yI2XQOvsTqcx48g/nUB53mF0rnp82/fF2bNxnprKomwyYn/HXFGBT2QHAjoNQKGsayCObfyK7BMlSLb/AoHkJH5J4dmXGDRnKe4BV28QnVRq2rg0fY5oS+C6M1DOzp6UlmZRM2omCycnjyaVkvfzC6K0NBGoXpY6A5VKVVXjqbEIgkB4eHi9BqUhFBcX06fPcHJyQqisnIhGc5xly3qzYcMahg2r6f9evPiVc8Zp7rmWMAyGIBYvnsysWTMcUuKjFTuyzYIlp+UJc1bHVTIj1BNerVKm46W1e116ecpVJd8dhc7VC1k+hT2ooNoRuXAcvfuVR/Zq9K4Edx1wxffnnTpI3PpPkaXZyFIEeUnfcvbAdvpNX4yyWiqHqayQ7OMxSGIadt0/QF6CzZJN2qEttB96R/0DtFIv152BmjdvNu+990+Mxu+xZ44b0ev/yZw5c5p03KeffogHHliEwfATEAkUoNfPZ/bsOZfNWWoqXn31TTIyemE2fw4IWCxgsYxi+vT7SU09VsNgx8YewZ5hX53eVFRUUlRUhK9v8ygBXI/YDCYKjzQuxPha01OpRiUkYvdGTMGeoRiLSvEZM8Y+ibT/FxTncnhkyXEhzq7+ETh5ulBRsBhZegnQAjtRKN8nvPezDhunMUiilaMbvkCybQDsRk60LqCy6BbSj2ylTd8L5WjK81NRqHohiTVD+2VxPCUZb1/LaV8XXHcGasmS5zlzZg4bNoSh1d6A2RzL2LFjeO21l5p03Bkz7iUrK4+lS/uiUHhhteZz993TefPNJU067qVYs2YDZvPHUCML6WYKCxeQkpJCZGRkVWtYWATFxUeA6ru1s6hUwjWRnGpqZFkmJyeHguxs1FotIeHhuLg0jz6caDJTciq5WcZuDB+N7MEDWx/AYH0WQXBFkpN4cWBnFHv+hCl241R9F1VhqZ3R2HgEQaDXlMeI+/ETSrM+RlC4oFTLdBk/Dxef0Mt30ASU5Z5Fln05b5zsKJBs95N9/IUaBkrv7ockHqN2eXeEQzh7N5+2pyMx2KwklhRSaTHipXchyt0LtaJpPCzXnYHSaDSsWfM127ZtY926dXTseCuzZ89u8uRbQRB4+ul/8dhjD5KWlkZAQECzv9idnJyB2uG5FkTRWCey78UXF3H33Y9hMAQDfYAUnJxm8sADC5ttB+goJEni4K5dWI4dI1ijwSyK7Nmzhy7jxxMcEnL5DhyM2cJVV0W9Fmjx4IM+Q/gl4zSF5jIG+vYiSrC/ZJO/+4W2ULWLOn8O5ZBxnT3oe/cTmCuKsVlMOHn6IwjNlxGjVGtBLqeO25ESVJqa7xUXnxDcA0MpyZqHLL4FeAA/olR9THifF6/ZnJuKfJOBrckJ9LZZ6STAcVlmvc6F8W07o2+CgI1mz4NyNLIs89Dcudxzyy1YvviCDU8/TdugIA4dOnT5m6+CsrIyjh9P5PTpFHx9fZvdOAEsXDgdJ6d/A5XnWmSUyv/Qs2dvAgJqhszedtttLF/+Aj4+d6DVeuHs3IeHHx7O0qXXRm6qKcnMzMSWkMDg0FAiAwLoFBzMAHd3ErZtq1MupJUL5Bor+ee+zZxKPUlQ9lk+id/Dm3F7OJVSM+m0l6fsMDefJImUZCWRlbCXsrw0tM7uzWqcwJ7wq3XRYa8qcJ5ilOqlhPase651w+0P4tc+DUEZiqB0x8nzKXre+TDOXo1Pn3GU0XcU+zKSmSLZuEOnp79WzyydE72MFRzOr53U7xiuux3UmjVr2P3tt5wyGjmvd/A9MHXiRE6mpzeJ+nVSUjLbtp0BQhEEFaJ4hIEDA+jRo/Nl721K5s+fS0zMIdaujUSpHIEgnMDXV+K77zbUe/2sWTOZOfNeioqKcHd3v27ylHJTUgh3dq5x5ubm5IRLcXHr+dol+Pj4AR6wmHjy3L+twITSQn5JS2JRhD0ISZUaiy28h0PGE20WTu/aTHm+H4KiB8jZZMZvIGr4UJw8/B0yxpUgCAI973iQg98uxWb5DIhAFrcT3G04/h3617lerXOmx233I1pnIdosqHUujQ7QCvV0Iub3BCQZkEHuE0yvXmPslXC/a54q0garDfFUGYM9dSiqPc4onYb3rPlEXGEw16W47gzUt598wj8rK6kuxnMn8EJZGYcOHaJPn7oS9leD0Whk+/YkfH2Hozknn2+zRRATs5Pw8KArCi93FEqlkv/971NOnnycgwcPEhw8j2HDhl3SSCsUisuW8rgaZFkmPz8frVbbpLvMwsJC0tLSiIyMRKlWY6tnp2ST5dboxItQajGTVF7Ko9Xa1MBzksic7LPUjJJ1DPnJ8ZTnd0DnNqXqhW4xRJF6cC0dR93SpFG4l8PZO5ihC9+iKDUBS2UpniGvoPfwu+Q9SrW2RoRfYwn1dCK92MAfvyfYG/oE07vfzVXl2i1WG4Vl5fi4u6JuosRpWZY5nZmLVRRp6++DPukseg93tMoL75AigwkvZx1tpzRc+Jb/rm3QZdedgbJaLOhqtQmAXqHAanVM2enS0lKys7Np06YNubm5yHJAlXECUKnUCEII2dm5zWqgztOhQ4cGCcw2Nfv27WPmzAc5ezYZWRa58cbRfP31h/j5XfqH3hisViuPzp/PNytXEq7VkmaxcO+MGdzSpQtBViuac7vCrMJCbL6+LeL/T0tElCWU1H1B6AGbXFtX7soxluaDLKNz96UoNRu1blwNQ6TWd8VQvAarqQKNvuEakE2BQqHEJ6L7NR2zPiPVs894lry2jOUffookKlGr4anHH+KRB+c51IgfP5HInHtnUpCVjUYhoHB145bx49iYls5tAYEIgoBNktiQkUmf6XNQ9HO85uN1Z6BunT6dDw4d4rZqpQd2AzkKxVXvniwWCwsWPMa3336DWu2HJBXx4IMLCAu7rZ6rRZT1JPH9XUlPT2f06IlUVCzHvqc1sH37vxkx4haOHt3X6B9Weno6KUeOYCotxTM8nKju3XF3d+fFp58m+bvvOGM242U2kwNMWrECl4ULsanV+MkyJlmm0tOTvqNHN+uqvCXjpdUToHdipaGiqryDDLwjKOjld/WBJeV5qcT9+CnGskIAdG5euAd2RZZrLyJFwIaiiaLE/grUNlIrDq3nmw+3YzAcACIxmk6w5PU7cff1Ydas+xrVd0VFBb+uW8fR7dtRqFT0uukmxt18M5IkMWniJJ4vLGQW9kX+hkoDM1Z9j/ruuzmWnk4IkChJtBk7jlG33IbQBLu4685AzZgxgx9XrqTvwYPcVVFBulbL90olK1auvOiZypkzZzhw4ABBQUEMHjz4oi6wxx5bzLffnsZkWovJpAJKeffdx5k2TYGXV3ucne0uK7PZgFKZQXDwoKZ6zL8cn3zyBRbLVOw5NQAuWK2vk5ramZiYGAYNavhnlXzqFOm//UZnDw9cnZzIOXWKfcnJDLj9dj7+8EMOGAxVIjkBwDKDgXtXruTZ06cpKChArVbj5+fXWo33MsyN7sNjR3azSZboIUn8pFBSoHPi+Tb1K4JIkkhl5nGys814BEddtIquzWJk//9eRbTMxq5UIWIoysdY+hkewRGotHMRBPuryVIRg1uAS4M1865Xzhup3TuOkvLNh1gM27DnWwJ0wmD4gH8vWUC3sQ1PBLZZrXy79Hm6pyQz3dsHq2Rh6wef8MKBODzbtqedyUz17NGbgYmiRIVXEN2nzKa0qJCBQSH4BQYRm2t04NNe4LozUGq1mvVbt/Lzzz+z/O23OZuURLS/P9nZ2YiiWOPMQZIk5s59mJUrV6FS3QicxN9fYMeOXwgNrZlzYTab+fzzTzGbPwUiAE+gAKNxDjt2rGDQoIEUF3sjCEpUqjxGj+7UbHk2jaW0tJS9e/fi4eFBv379mmRXcfLkWSyW2kZIALqSmpraYAMliiKn9+5lSEAATudSB9r4+yPm5JAYF0ep0Ujto9p2QE5REXq9vs7/11YuTltXT94ZMJb1aUl8kp+FWbTR3smVXGMlXaip6pCZmsJ3D0/FWOZEhhCJLP6XkB7D6ThqWp3vU27iPiRbBOALjMP+GtqBLLVD43QEc/krQDSQjc49n/DeI6/NAzuAioIMjKV5uPqGo3NzbN5TqKcTaQVlWAwFQO0ArO7k56Xy1YG0BveXl3QUn7hjRHsHUFxuP6Ptq/Hk3R27+T2tiJ71KNdHmU18F38KS4+JQACk2yC94WM2luvOQIH9YO/9N95AfeQI7xqNmLOz+c/DD7Nz0ya++v77quu++OJLVq06iMmUArgCMkbjK9x5533s27e1Rp/l5eWIojMwFHtFTrDnOIwgN/c9+vcPJy8vHy8vT6Kibmz2oocN5d13P+Cpp55Brb4BScrB21vgt9/WERUV5dBxhg7tzYYNGzEYqq/JjIjiTnr1+neD+zEajaiNRpy8amq1+bq5kZGfzw1RUaxPTKxR2O0HYHDfvlc1/78rBSYj2zKTeUiSGCXL7CvIYklRLi7eA6oO62VZ5sl5C6gsegjk82EVxWTGD8IzZC8BnQbW6LO8IB1ZCgEe4UIyaweQDwCn8ItyQrIm4h4UhXvQwGYPM28IVlMlh9e8R1lOOgplZyTbBwRGD6TzTTMQHOieDPNx47RrONbyLUD1M5+NaDwiiTl/VtUAKpP2cXdRJXmW4hrtfhUV7HO3sV4GI/ZzR7A7W/+n0lIu+TVqnKvhujRQGzdupDA+ngNGI+e/GuMqK+mwYQNHjhypKrvx/vtfYzA8B1UxfwKi+ARxccvIzMwkOPhCRVAPDw+cnAIpKyvkgoECOI6HRzS//25DEHyR5UxEUdHsIeYNISYmhsWLX8VoPITRGAnIVFZ+xJgxk0hJSXDoTmrGjHt5/fV3sVofwWqdBxSj17/A+PFjGhXAodPpsKjVWKoFPACUVFbi1LYtr773Hv+YOJGzRiP9ZJkdSiXLdDo2vd0qM3MlrD4dzxJRrKoNMARoJ4m8uC+Oaefakk4nk5uZC/JD1e70RLQ+R9rhZXUMlNbZAwQdyNVf3CLgR2nOWWzmG4ACijNjiRru1qwh5g3l2KavKc3uiizuRbKpgDKyT4zDxXczbfqOd+hYnUffSfz66Ui2d4D+wO8oVE/QZfSDeDdCSb7I35+8swpctDXNQJ5ZSWT7LmTl9mTImSM8bzWjBd5UaSj2DaVr9wFXbXSPN/C6lr80uQJ279jB5IoKqn+EeuBmUWT37t1VbRUVldhdddVRo1Q6U1lZWaNVpVJx7723oVa/CmwBkoHvUSq3c+ONUwgJ6UpwcCcCAoYRE5NLYWFhkzxbfeTl5fHccy8xZMjNzJq1kKNHjzbovvff/wKj8TEu+LIFZHkBRUV2dXNH4urqyp9/7mLWLPD3v5WIiAd56aXxrFr1RaP6UalUhPTsyeGsLIxme+HGgrIyTlosRHbpwsiRI9m0ezdHJk9mUceOnJ06lV0HD9KrVy+HPs/fhYSyIu6q1TYROFFWjsliD2ioNBhQKNyA2i8tT2z1uIm8wjqj0haCsAnIPPffehDMOHvOQO8+Cr37VGTpHlL2xdCYkkBXS3H6CWJ//JAD37xFyv6fsZkvf7YiWs3kJe07pxxx/mXvhmR7g9RDOx0+R/+oPvS8Yz7uQW+j1g/GM/QTek99tNH1pdyD2nPYyY248mJEWcIiiewpK+SsZwAuPiFE3fooppEzeTCgLbN9w8kYMoXO01506I7wclyXOyj/oCCO6XRgMtVoT9ZoGOZ/YTV2++03sWzZp5jNA7mgV/cb7u5a2rVrV6ffBQumkJf3E9u2/YDBYCEszI+IiN4MHXpBGVylUqNShZGWlo33NdDeysjI4IYbBlFePg6zeS579ybw3XcjWb36v4wff+mVW35+CbJcuwibgCAEUFxcXO89V4O/vz8fffR/fPRR7bIkjSO6e3dOKpXsPHwYuaAAna8vXUeOrPq8e/bsyddr1jhiyn97PFUazlhMNYpupANOCiUalf1F1SW6E0plCRADnN8tyShUHxHQsWudPl392+ATeZCi1CNYDAUIggKlJg+lKhetS8+q69T6aEzlWswVxehcver042jSDm/l5Pb1SLYngDBKs1eQEfsSA2a+gEqrv+h99rLxCqB2Xl8gNnNFk8zVu003vNt0u6o+lGotbsPvYc2R3/gx+zQyCsTwznh1H2V3qwoQ0mMU9BjloFk3nuvSQN19zz10eeEFfgVuOte2AjimVjNx4sSq65566p+sWTOM7OxbMBhuQ61ORK3+iq+++rbeCK9Vq77np58+xmy+BfDmzJlfcHUdREDAA7WulFEork348gsvvEpJyVRsttcBEMVbMRj6Mnfug6Snj7tkpNrkyaPZs2cFBsM/uGCgU7BYDjNw4MCL3teUSJJEfn4+kiTh4+NTb+SlQqGgU7dudOjSBVEUrxvFi5bI6NB2PJRygp8lET/syo4LFApu7xBR9d1Sq9U8/Z83ePahWxCts0Fuj1K9Ep17DmG9nq7Tp81soCTjIFbjaZDbIctGbKZjOAc9hkJZK7CoCWu41Zjo9wwmAAAgAElEQVSTxcTJ7auQbAcA+/mrZJuEqfx20o5sJbL/xQslqnUu6N0CMRT/BNVPP4Uv8Ym8tnlT1bEYyjCW5qHSOuPkGVDv56h18cR3yBQkmxUE6q1v1ZxclwbK39+fNRs2MGvqVJQVFVhkGTd/f35dtw69/sJKyMPDg7i4GP73v2/Ytm0PbdoEsWDB/hoq3+dJSkrirbeWYzYfxV7p04jVuoi4uJEkJMTSpYtd7sVqtSCKqYSG9qzTR1OwadM2bLbVtVpHUVxcSmZm5iWj1mbMuJePPvqapKTxGAz3Igg56PVv8+qrSxqt8pCVlcXq1asxGAyMHz+e7t0b/8MsLCxk06ZDlJW5AEq02jhGjepMWFj9z6BQKFpDxZuYCaHtKTEZaZ+dQrigIFWSGOQbzMJeNd1Jg0eNY/Jrn7Htyy9RW4/gGxlNQPRclKq6QsNn9qzHXDkSWfocKMR+/rSc8rwduPjedCHE3BCLk6cNjfOV14FqKGU5ySiU7ZFs1YODBCTbTPKSllzSQAmCQOeb7uHw6tmItoMg90RQ/opKs4GoYS82ah6yLFOcfpzi9EQ0zu4EdBzQ6DL3siyTeXQfuYlZIHQA+SxOXgdoN3gkal39kcWKFlqZ97o0UADDhg0jKTOThIQEVCoVnTp1qncF4ezszPz585g/f94l+9uwYQOSNBk47yJUA27YbDeyb9/HeHjMBVQIQg5Dh4bj5dX0LgkAT08vsrIygequlHIkyYir66Uz7/V6PXv3buWrr77mhx/W4OvrwQMPrGpUThLAmjU/cO+985Dl27Ba3Xn55fHMnj2Nd999o8GrX5vNxq+//olC0YuQELtDyWisYNOmPUyb5vmXCdm/3lAIAtOjujMpohPZxgr8dM64a7So6zmGcA8MJ2DgHZct+Z5zMhZZXIt9137eefgCkhiAoXgJCmUvBCEftT6JNv1uvCY7KLXOBUnKoa5ieTpap8sbCK+waAbc9xKpB7dQWbQLj+AQwnotRevc8IWeJNo4vOZdSjLyEa2TUaqTOLl9Eb2nPo5HUPsG91OSeYqc4yI6t+cRFFpkWcZQvIPUg3toN2Rsg/tpCVy3BgrsK+xu3a7OT3serVaLUllZp12tFujbN4Tx470RRRF//0FXXEH3Snj00dk88sizGAy9sf/YrWg0TzJmzHg8PC6/8tTr9SxYMJ8FC+Zf0filpaXMmDEXo3E7YN9FGo3P8eWXfbj99gkMHz68xvUWi4W4uBMkJNjVj6OjA+nRoxMFBQVUVHhWGSf73FwoKgolPT2TTp2aX6rp74yLWkN7tWMWXXY3Uu2zGREEMx1GdMZqLEClccLN/7ZrtrJ38Q1D7+5CZeFrID+F3UiloFC/Rljv2Q3qw9krkOix917xHDLitlOcoUSyHgM0iFaAH4hdt4hhC9+sY6gr8tPJPHqUioIitC4uBHZuj1dYNPnJZ1Fqb0NQ2FNdBEFA6zKU0uztWE0VF91FtURa/SMNZPLkycjyBiC2WutJlMpVTJs2jfDwcCIjI6+pcQKYPXsW8+aNQqdrj7v7CPT6cPr1O8tXX31wTcb/7bffUCoHct442fHEYJjHypU/1LhWlmU2b97LoUMq3N1H4O4+giNHNGzcGIPVakUQ6r6MBEGNxQGF8FppOYT06I9C9SJgrmoThNfxDIrGPaAtPhHd8QiOuqZuJ0EQ6HXnwzh7fY1SHYFKOwCFqjvth45pdHTclZIZfwjJ+i+gult0MjYTVOTXTIatLMri5O+HMJRMRuv6OqJ1Hil78yhIjke02FAoagd1KAEtkvjX+i1d1zsoR+Lv78/LLz/L4sUDUSq7oVQGI4rbef/9ZURERDTbvARB4J13XmPx4kXEx8cTGhraDMKwDQsDzs3NJT1dRWjohRyx4OBo0tNL6dpVRBBysVrNqM8pQIuiiChmEBR0bV4QrVwbwnqOITfxVcrz/BGEoQjKdDT6ErpOfPLyNzchendfBs15mfK8s1iNFbgH3n9tJZYu5cmstXvKOXEMQTEJjVM0ACptCIJiBlnH3sI/KpyMuP0oNeFVuy6bORmtixmNU/PXqWsMrQaqAYiiyLRps/n55y2oVFOR5eOoVHvZtm0T/fr1a+7pAeDn58eoUfZwUEmS2LZtG0ePHqV9+/bcdNNNqJpIjn/s2LGI4nzgCHDDudZinJw+5u67a+Y4lZeXo1DUdRMJghc2m41hw9qyY8dulMo2CIISqzWNnj09Wus1XUdUFGRwcOXriLZIBKEnsrwRj8BIet21tEVEkAmCgJv/hQWn1VRB7sn9iBYTPpE9cPYOvsTdV0dwt95UFvwH0TqOC7uoH1DpqFPu3lBcjkpbU9RLqfbBVK7GI6QdJZm7qSj8EoWyG7KUh0L5B236No2MWVPSaqAawKeffsYvvyRjNJ7mvPCH2bycuXMfIz4+pnknV4vS0lKGDRtPcnIlFstwtNof8PFZzJ49WwgMbHxFz8vh5ubG119/xvTpI4FbsVo9UKu/Y86c6QwdOrTGta6urkhScp0+ZLkIV9dwgoKC8PPz5uzZTERRIiyso0NLcbTSvMiyzJG1H2AxvAicP/OspCRrKNnH9xDcdXjzTa4eClLiOLJ2OQIjkCVvTu18mdAbhtFhxNQmedGHdLuR/FMJFGd0QrRNRqFKQhB202PS43XG03u4UpabilJ1YcEnWgtQqq1onN2JGj6ekqzTlOf/gdZZj1fY6L/c7glaDVSD+PjjlRgMi7mgSgWyfD+nTy/h7NmztGnTptnmVpvFi1/kxIkoLJYvAAGLBYzGxcyb9xg//7yqSca8/fbJDBw4oFqY+aZ6g1P8/PwICUkkI+MY/v72qKS8vGSCgkxVJei9vLyuWQRkK9eWyqIszOWVwNxqrc5I1mdIP7K0RRko0Womdt37SNZfgMHnWl8nI7Ynvu26Nsm5lEKpouddj1KcfoLijBNonQPx7/B2vWHmgZ2iKc1eh8WoR61rj2jNwWr8nrBeUfbSJAolXmGd8Arr5PB5XktaDVQDMJnMQG1ftBKFQofZbK7vlmZj5crvsFh2Ud2hbbM9xaZNAdhstiZz9QUGBvLwww9f8hqFQsHYsQOIjT1BQoJdjLdHj0B69BjQms/0N0CyWQEddWOznM/9reVQmJqAXTF8cLVWL0TrQ2Qd3dJkgROCIOAVFo1XWPQlr3P2DiZquEhm/GoqC4vRODsT0bc9Xm1avgZoY2g1UA1g2rSJLF36PibTMC68+H/B01PvcNXvq0WSROr+b1UhyxKS5LhKqFeKVqulX78e9OvX4/IXt3Jd4eobilJtQrRuAUafa5VQqN4lsHML+z5IEvW/HjUt4ncE4OobRseRYcjXSG2jOWhdtjaAxx57mKioTFxchgJvo9PNx9l5FitXftrivhi33XYbavU7NdoUiuUMGzYOjaZuVv+VUFRUxOnTp0lPT8dmqxm2Gh8fz9KlS3nnnXfIzMx0yHitXB8ICiXdJs5BoboLhXI+8BZKdX9cfNIJ6+n4cuFXg1d4F2TpMBBXrbUSpfp9gqIdIzws2ayUZJ4i/8wRKouyaojiSqKVnBMxnP5jNTmJe5HEi+8wW9o7yJG07qAagIuLCwcP/s66devYsSOGsLAI7rsvvkmCDq6Wt95awh9/3Eh+/kgqKkbg4nIAJ6c4Pv10+1X3Lcsye/YcIj6+DIUiAEkqxt39BDff3A83NzcWL1rE1x9/zFSLhXKVim7PPMMHn3/OlH/8wwFP1sr1gHebrgyZ9xqZR3diKt+Bd5sh+LXvg0LZsl5FKq2eLhPmkLBhGLI0BVnyRaFegV9UB3zaXr2MmamskFO7dmA1tEOWAxGEeDxDj9Gm3wishjKOfv007UwVjLeY2KrWcdjJja73vmIvVfI3omV9K1o4PXr0xNu7Da6u2hqafi0JX19fTpw4xNq1azlyJJ6OHW9lypSVDkkgTktLIzbWQmjo8Kozo8LCLLZvP4yPj4rvP/mEY0ajvYCJKPIwMHT2bMbedFODVC1a+XsgSTbUel9kyQ1BqUMSbS3OQAEEdhqAR1A7so/vwWYpxLftPDyCo656xyLLMikHYrCZb0fn1utc21iKUr/ELeA4hUc2M6O8iLdkuytxqdXEI+VWft7yJVG3PXbVz/VXouV9K64xDfHfms1mfvllD7m5Xri4tMFiqSQ29iATJkQTEtJ0eRFXilar5R//+Af/cPDOJTExCze3yBoBDd7eQWRmnmDLhq+Ye944naMrMFSlYuPGjQ6fSystj4bUbSrOTOLMnuMgjEah8qEkK46C5F/pMOImVJqWt+jTu/sSOeA2h/ZpNZZjKJLQul7YiQmCEpVuGAUp35F1+k+elmuecz0tiXycdICWdeLd9Pxtz6A2b95Mp059USqV+PiE8cYbb1/0B3bq1Bny8nwJDe2Bp6c//v6RuLn1Ydeu4y3mwPRyyLJMUVERBQUFiKJ4FT3VNeaybPeD1/fpyVzfPvK/Oxarlaeee4UxXTvy+bS+nFn7NqXZdXPdwB7Ak/ZnLCrdfHSug9HoO6J3n4KxtCcFZ65NCXFHYDMbqchPx1R2pUVJZeqXjVBUibLU/i3Z7/j7/Y7+lgZq9+7dTJo0g8TEZ5FlC4WFP/PSS6t4/vkl9V6fklKAu3vNnZKLiwdlZco6lXdbIiUlJaxevY3vvjvKmjUn+eabrWRnZze6n6ioQMrKztQwysXFOfj7K5g2cyaf6vVU/8nGA3/YbIwbN+7qHwLIz88nPz/fIX214hhmPLOcTz8/jdGwHzBgyn+Egytfp7K47vfLXFGMzeKBShNUo12lvYGSzIJrNOMrR5ZlchIPEf/zL5z8PZtjm/ZyaufGBlXdrY5a74aTJ1gN8dX6lrCZfscnIpiQ9n14pVrVWhl4RaEkMKqPQ55DEm0YywoQrS0rRaY+/pYuvueeewOj8VXsxasBumMwfM877/TmmWf+hU6nq3G9i4uGoqKaX0L7S9rS4ovliaLIr78ewGbrRkiIPRm2srKUDRv2MW2aO05ODdcaCw8Po1u3PBISdiEIAYABF5d8bryxL56enky7/366fPABd9lslCuV/CgIfPzll1d9/nTy5Enm3303sedK2Xfr3JlPVq6kY8eOV9VvK1dHRn4RG3buw2RJB84rZN+HaEvi7P7f6DxuRo3rlWotyBXIsq2q5hOAJJai1jsmwrQpKctJJiO2FK3rMyiULsiyRHnOZlIP/UHbgaMv38E5BEGgTd8BJO1chakszl7VWojHM9SKV/hIXP3DWZFzhj+MZdxoMbNNoyXNyZ0uo2Zd9TNkHt5M2s6VqEURIzLB3UfQZuRMe3JvC+QvZaAyMjIoKiqiY8eOVxUynZiYyIXS1OdpA+g5deoUoqjAZLISFuZHYGAgnTqFceJEIq6u3mg0OvtKKieRDh086xgzRyNJEklJyRw7lonNJtGhgz+dOrVv8PPn5uZSWupWZZwAnJ3dKSkJIS0tg44dG+7VVigUDB3al+joAoqKitBqvQkK6lplpF95802m3XcfGzZsQK/X88odd9SIdDSbzcTGxuLq6nrR+ly1MRgMjB48mH8VFrLlnAv209hYRg8eTGJq6jVXj78eECWJdEM5eqUKf/2Vf37JWbloNdGYLLWr4A6mLHcLxRknKcvJRa1T4xUehc7VG49gZ0oytqB1G4sgKJBspUi2Tfi2bfrTFXNFMdkn4ijPLULjpMO/Y3vcA9s12AWdl3QGpeaWqqq/gqBA6zaKksw/sJoqG1VYUO/uS+ebJlKWcwarORUnj/Y4e4fYS2M4e9Bz3v+Rl/Qn64oycfEO4YZ2vWoEkpjKizCW5uHsFYTGya1BY+Yk7qV4+wp228x0A7KAqXHbOSsoiBx1X4Pnfi35Sxio/Px8Ztx+OwcOHsRPraZYoeDN5cu5e/r0K+qvc+docnN3A9VVv5ORJAO7dmWg0bRHqXTl8OGzdO6cwdChfRg5soI9e37HanUFDLRr58KgQY3Lh8jJyeHIkSOEhYXRuXPDMr7tYd0yXl7dUShUxMSkkJoaw4QJQ1AqL7/qsZexqGtEBUF7xSoYPj4++Pj41Pu3zp071/tsa1av5sE5cwgEimw2/MLCWPXLL7Rt2xawG6Ls7GyCgoJqREj+8MMPdDOZeKja+eBCWWaT2cyaNWuYMWNG7aFauQQHCrL4/MRhnGWJUlkixMmNB7v2x1fXeNXuCB8vTOYEoByoXhxzGxaLyLHfixGUA5ClElLif8enZxTadr2gMobS/L0geIMiA8+OkZRq/CgtNlT1cKmih7IsU5Z7BquhHPegdg2qb2SuLOHE1q1I1ptQ6TtjLC3g9K6fCO9twLddw6o/28xWFMqaRUDtJWL0SDYL0Dhjr1Rr8QytK0WUfv5z8O+Gxr8bFiCzzAJYkEQrxds/piI1ljZKFUdFGx4dhuAxeDqCYD+xsRlKkEURlYtXDeObv2s1n50zTgBBwEqbhagjW1H2mNwixHpr85cwUFNvuYWehw+zzmpFazJxBLh5wQLatm9P//79G93fyy8/yb59kzAY3IGbgaPo9fMZPXoa/v7D0J1bCclyGAkJMbRvn03Hju1p27YNpaWl6HS6RlV4lWWZRx55kk8++RSdrg9WayJdu7Znw4bv8fb2vuh9JSUlHD1aSkjIiKrIOSenbqSn7yMrK+uS5dzPYzckidhsVlTn6utIkoQoZhIYeG10uhISElg4Ywa/Go30xl6zdPmpU9wyYgTxycm8/OyzLH/3XdwVCspkmUcee4znXn4ZQRBIS0uji8FQp8+ulZWkpqZek/lfL6RXlvHRsYOsl0QGAVbg9YpS3oj9gzf6jW5UMEtKKkRwkBEhAfyeNQmj+V0gFFiJQv0Z7j6P4BH8YFWfVtNApJz/Y9Dt41GM6YOhOB+LsRIXb39U2prRe3/sSCC92FCvkTKW5vPnd+9gLrchKEKQxHdpO2gykQNuvuR885ISEK0j0bnaBYyVKm8U6jlkHn0D74joBr2cPYJ9yTx6GJU2rKrNZk5FozeiaUTl3EuRXmxAkmHwjfVLKf35xev0So/nO9GKi2ilELjpdAyqHl0J6z+aA+8+RUHqKVQI6Dx96PPgK/i1t1fbXvtNaY262wAhgEYBPXuF4uRR/6KzKTj+ccOua/EG6uTJkyQePcomq5UvgfeBHCDUYODVF17gp82bG93nwIEDWb9+JYsWvcCxY/fg6xvGwoX34e19Y5VxAruv2MkplLNncwkODkatVl9053ApvvjiSz7/fDtm82nMZm/AxuHD/+See+azceOai95XWlqKQuFbR6dOrfajoKC0QQbK2dmZgQPD+OOPP9BqI1EolBgMqXTr5nTNylh88dFH3G+x0PvcvxXAw5LEl8XFPLRwIbHffEOc0UgokArc+c47uHt58ciiRfTu3ZsnnZxYWlFRFdEjARudnXmlj2MOjf8ubMs8w/2SSCSwAPgZ0CNTZqwgoaSArp6N+z6kpMLLQ6L5KPY0a84Mo8JYTvQNg4mImkVJ4U24e1bfbUSRnxvGqEAIjQgDwi7WbVWQW8zvCTWMlCzLHFr9LobiWSA/ab+IdJJjhuAWEIpPxMV3QhV5Jag0Nc8slSovrEYPLIZydK6XFyj2bduVorSNGEtNKNVdkGwFCIqtRPTrWbV7uRqqjNPwLszoW/fzEUWR72auZ73VXHXi5w0st5iYunUVOb+vYV5ONoskCRWwNi+T+a89wBfb9+Pt60/yDT3ZsPt3HqzW5wHA2dWVeSO7N8gj4yg+a+B1Ld5A5eTkEKFW8wr2H9RyoC2wBnh261ZOnTp1RXp4I0eOJC5uZNW/CwoKWL36RJ3rbDYLOt3VfUzLln2OwfAy9q8TgAqr9RV27AihqKjoourdTk5OyPKFSpqFhYXk5+ejUBTi5tZwFYtu3Trh7+9FcrL9HCsyMpzg4OBrFv6dn5lJl1qh7QLQRhBY/e23bDcYOG9qw4EPDQbueuMNHlm0iNGjR/Nap05MOXqUx00mBOAtnQ7XDh0YO3bsNZn/9UKZyUAYMAx7eNBu7M65p4Bvk+Lo2ndUo/vMSFdws3cUD/WMou2UmxF6jWH5x9+ze58ZX+cL56SyLGPWWOgd7k1osOslegT6hfHV/jQGDu9CzO/28PNQTycqCtIxlZaD/AQXwrRDkaxPk/rnN5c0UFo3J4xlWai0ISDLWIwVyJIRKGlwUUKVVk/HkRMoSjtBWe4mtM5afCKGoXO7uBekodQwTv3C6F3PZ2Q0GjHbrATVam8DFJWW0Emj4alqEbZ3AJslkaO/reXxJ57glTdeZeLIkYhGI+NkmTjgCScnXn79NfqFtcxE+hZvoHr06EGCyUQCcAw4H+z9CFAsy7y1ZAkff/31VY/j7e2Nj4+ZwsJMvM8VJTObjYhiCpGRvS9z96UpKSkGahsUZxQKJ8rKyi5qoHx8fAgOFklJiWfFio3ExsaiUrkiir+RlTWNN998pcFGxt/fH39//6t6jitl2IQJfLdlC/dVVla9VvLh/9m7z8CoqrSB4/97p8+kT3ohhUBoEiB0VBBUUFBUVCxYV9d11dVdX/ta1rW7uq7u6trb7irYxYIiTZAmVVpCSUjvdTJ95t73QxCS0BISJhM4v2+c3HvnmZCZ595znnMOy9xuGtxu2q/bPAgormkpO5Zlma+WLOFvTz3FDe+/j6qqXHzVVbx1331iBfROyoyK46XaSgaj8rdW7V8CaU47e2z19A2NPNzpR1RQCMz9ilSXm8mnD+WHJd/gdgxEr28ZwK+oXEFGqkpSVBiqo+mI18qJBHVUEtK+yrKVS1u6/EJcdpBiOHh2TAJex5Gne8T1H0h90dc46o3Ul9hQFQ/wAxr9CpqrhxGR1LGbXI3OQEzfYcT07dDhHdI6OV09KomcSPWQvyMjMDQjg8937WJWq/YPgH6pqQwsLj7onCEuF7m7dqI6mhiW1Zf5387nb48/yUubNtMnOZnn7/4/zjn7rKP+n/QUqSOzv381cuRIdd26dccxnEO7/ZZb+Orll2k//W8lcOeAAazacfCTz7FoampiwYK11NUZAT1abR2TJvUnMzOjS9f93e9u5803Nfh8z7dq/YHExFsoLt5xxC9al8vF+edfzZIlPny+q4EYIBaz+TJeeOFmbrjhN9TW1tLc3ExYWNhBya6wsIjNm/fS2OgiNTWSYcOyCAvrWNVPd3E6nUwaNYqk/Hx+43RSCzxlsXDxLbfw/Vdfcc/27VzY6vh5wEvZ2SzftCmgcXaEJEnrVVXt0h1LZlik+tzIyd0VUoe5/D5uW/ENDys+bmn3szmyhqj+2UxJSOvSa6SnQurMs/h2Qx7vfV+EovRDpZ6UmDrunzOBuMiOjdVIOWezvlGLJGt4d20RK5ZsBZ+bXe/fjd/7E+y/rVGRtRfT91QLGWPPP9IlqS3cxroP3wT1NCAaGAvo0ehvZOLv/46s0WKrKkRVFUKiU9pU5fncTipyN1BXWIGkkYnpm0xsv+FdXqKpfXIaaZVR1nx12ON//CWXax56gTu8XsYoKos0Mm/q9Tz/p+u5929vsMvt4deSKBWYYjRwxa1XcfmU9lXLPSvknOs79DnqFQmqpqaGvomJFHq9tH4QfUmSWH3++fz388+77bVUVaWmpgafz4fVau2WFcDLy8sZPnwCjY0TcbnOQ6PZisHwTz7++F3OOeecI57rdDqJjIzH7c4DojiwFfQP9Ot3Dy+88AJ79yrIciSKUkv//kYmTRqFVqslN3cXP/xQRmTkYIzGEOrrK9Bo8rj44vGEhh6lm6WbNTc38+orr7Dgo48ICQvjmltvZebMmSxatIgrZ87kUYeDCbR0Oz1sMjHvm2+YNGlSQGPsiN6coADm5m9HLszjP63WKlCBQRotc04Zx5BOjkMdSvq+ncibPV6Kbc2YtFr6hIUgd6JLOXXmWegmnIek1bO+XtqfpBpzl1O5cj6K704gDVn7HsbQXMZd+9BBxRbtFW9aSO4PNSi+t2nZfLSlMEKju4g+I2XstTKKbyCgRZK3kjpyENa0wSh+H7mLvsJRPxK9eTyq6sXr+IHIPqVkjOtccUmbeNo/OYX7UNd/z565h09QALvqG/lwy06KG5rIjI7kilOySA61cP/iVTSXVvGg308I8JIssynEzDszp6AP4PhSR2S/82mHPkdB38UHLV1dsy+9lKs//ZSXnU6SgO+Ax0wm5t9/f7e+liRJ3V48kJCQwLZtP/Pyy6+yePF7ZGamcMcdSzpUam6z2QANEEfb5VHSqaysprAwkpSUloofVVXJzd1IbOxOBg/OYtWqPcTHT8Sw74MbF5fG3r12vv9+KdnZA0lISAjYPKKQkBDuvOsu7rzrrjbtZ555Jp//8APPPvww/9i+nUFDhjD/L39hzJgxAYnrZDMtuS9/KtnNq34f1wEO4BFJAoOJwd1UxVWwv7hSh5FIVKCwoZMX+WIhfY0GyDmbnEg97CsaWClBn7QMitb9gLu5mZh+WSRnX9qhdfw89iYUX1+gbQ+C35tC1c51hMY9jd7cMsLj99Wx9+fnCYlOxNFQhbM+FVP4gSc0jfYKavIfQW9eTGhMMmFxGcjajpdpF9c7GD+p5XN7zeg+5ESqKGtaklPBUYpTtYQzJ3VUy4At4K2Dgjq4Pn0M3+j3cFtZAV5FYURMMvenZVFaElzJqTN6RYICeOnNN3kgIoJT3noLr89HRlISb//rX4wePbqnQ+sQq9XKgw/ez4MPdu68mJgYYmPjKS5eCBzYM0eSPiAzcyRxcf1btUnExWWxZctK0tNT8HgM+5MTQHl5BZs31/DLLzXs2JGPoixnxozhjBjRs5vFjRs3jk+//75HYzhZhOsNPDhiIq/mrueO5kYkYExUPPcNGBF0aybWbtyBFZDGtC0hj0jsR8T5/Tp9vag+g9Do3sHv/Stg2NfqRtbMQ6uf02YJJo02Cq8yjobyPfhcXpAPfEZURZveK28AACAASURBVKGpohSXrQ9F62uQNVvRmRYzaOpFmCM6N8577dhUVKXjvVhHopFlzkvpx3kpnf/dBKtek6AMBgN/++c/efLvf8fhcBAWFhZ0H6jjQZIkXnnlWS699CqczrtR1WHodAswmd7nggueOGiJElnW4PP5MRqNaDRuvF4POp0et9vN5s1l+P1G7HYXpaVW/P5Ynn12GTff7OS008aeFL9PAdJCwnl45GScPh8aSQq67p/jJSJ5AJEpCdQVn4Hi/T8ANLpnCYmJQWs8RKGSpEdVVAyhIaAWARMAcDbV4moOQVF24nFEIMlTcTYVsuGjuQy7aBYh1uDb4aC36nVlUDqdjvDw8JPqy3T69On8+OM3XHLJVoYNe4ybblLYsmUtp502hOrqvW2Ora4uYODAeHQ6HTk5yZSVbcTtdlJfX09zs4/i4hVER08jOjqHuLhhhIXNYNGiEioqKnrmzQk9xqTVnjTJCVpu9oZffBsDz8wmPOkpIpKeYsCZwxg+63YkaTOKr3H/sariQmUNYXF9iEzqj870Cy7bKlTVh6uxHlX5EhQjGv2TaPUXoTPejtd5DbuXL0dRurJbgNBar3mCOtnl5OQwd+7bbdqsVisVFSspLm5Ap4vE660hPr6ZoUNbKnaysweh0eSxYcMyyssrgSaSk7OwWtP3X0OWtRiNaezZUxaUOwQLQneSZQ3J2VNIzp7Spr3PiCyK1j+HygQktMAqEgdZ93fZZZ1xJsUbl9JY/il+XyWyNgpJdxOyfGAZMVnTB587Dkd9hXiK6iYiQfViFouFWbMmUVpaSkODnaioeJKSkvbPCJdlmaFDBzJkSBbNzc28/voCcnMPfKC8XjeyXLtvtfHgnAchCIEQ0/cUQmMSaSjLR1UUwhNGYI48sMCyISSSzNOmovi8VO7awO4f9+L3tVqR3VeF3qxHkoNvPbveLOAJqmUl8ApqysvRGQykpKWJFam7QKfTkZaWdsRjZFkmLCyMCy4YxeOPf0RJSRRGYySyXM+QIXEoyh4yMjIDE7DQbRw+LzsaanB4XFhNIfQLt6ITk5ePmTHMSvxRVoWQtTri+o2gtmAnVTvfxavGIUluNDobhlAfWn055shxAYr4xBfQBKUoCmuXLcO3YweJOh0uv58VK1cydMYM0b3UQYqiUFZWRklJDUajjoyMlA5PvO3bN52HH57N3LnLcLkSCA+PQVXzyMmJFL//XqbSaWdJ/jZG+rwMkmCbAl+aLZybMQSTVnSMdITHaaOuKA+3zYHFGklkcv+WPauOQtbqGDTtSgyh86ne/TiyPAat0YMkbyZj3Nig3VupNwroX3JJSQnq9u1MSEnZX+SQZLezeuFCYq+8skuLFfr9fj788EM+eecdAC65/npmz559Qi2HoygKixevJi8PTKZkfD4Xa9eu5pxzBpGSktyha6SmpnL77bMpKyvD6/USE5NKZOSxLW8j9Jw1JXuYrfgZZmyZRjAW+NTZzMaaMsbHH2Eh1g6oczv5tng3hY21xJhDmdanHymWwK4+crw56ivYuXQ5Pu94ZE0i1Xt2UJn7NVmTp3VobT5Zlsk89XySTqmiuaYEjVZPeMKMo04WFjonoAmqMj+fPhZLmwq8cIsFS0MDDQ0NR9x64khUVeWaSy9l13ff8Qe7HRV4ftUqFnz2Ge/MnXvCVPwVFxeTlyeTkjJm/3tyOBJYtOgnrroqocMJ3mAwkJ6efvQDhaDk9HlxOmwMNbTd52u8Ts/L9dXQhQRV7mzmoXVLuNTv52pVYX1TPQ9XlXDH0HEMjYztauhBQVVVitb/jKpegSns1w0ocnA1fUVF3maSh3asi06SJMwRcZ2e+yR0XEAfLzQ6Hf5Wq+3+yqeqXXrSWblyJau/+45ldjtXAnOAH+12ln/zDWvXrj32gINMQUEVFktKm4RrNoficoVSX1/fg5EJgSRLEn4JfO3a3ara5S0TPtmzjVt9Xv6pKswAHkblbcXPe7kb6cyyaMHM73Fhr3OjM7Xdc0lnGk19sZhuEUwCmqCSMjPZ43Lh8Xr3t5XW1qLExu6rJDs2S5YsYZbTSev7SRNwkdvN4sWLO3QNu91OQUEBRUVFeDyeY47leDIYtPh83oPaVdUT0L1cAm3jxo1cecEFDEtPZ/aMGSfUTcexMGi0xIRHs8Tt2p80fKrK914vqdb4o5x9ZFvqq5jTrm0GUOF2YvMd/XOhqCqFzY1srquiyN6EEoRJTdJokCQ/qG3fj6o40OhO3PE7RVX5vjSfh9b+wL2rv+PDPduwH+L7JJgE9H8jLi6O+tNPZ/FPPxGrqrhUFXtkJGPOOvYFFwGioqLINRjA6WzTXqLX078D3Ya7cnPJX7aMeEXBL0ls1+sZdu65xMcf+sPu9Xqx2WxERkYGtPswMzOZDRt+wetNQLdvMLe6upj4eLqU4I+VoihUVFRgt9sJDw8nJiam238fP/30ExecfTYPOJ3cqaqsKixkxuLFfDB/PlOmTDn6BU5Q45My+N7jZovDRhKQB4RZ45gU1bXuplCtjlKfl9br99fRsqCsQT7y14XT5+O7vdsJt9vIAHYDGyxhTEsfiPEwq347fV4kSTrsz48HjVZPZJ8Y6goXYgidjiRJqIoHj3MBSUNTAxZHa/UNDWzYuJ2ysiqMnnDC9Ucv1uis13aso7a6jCcVP+HAv4p38Uh1KY+NmoIhSG9wA367MGDIEPpkZFBbW4terycm5uAdYztr9uzZPHT33SwEztrX9h2wSJZ55ZJLjnhufX09hUuWMCkuDoOuZQ5DQ3Mzqxcs4Mw5c9C2qojy+Xw8fO+9/PuVV/D5fMRZrTz+979zyezZXYq/o2JiYpgypQ/Lly/B77eiqi5iYjyceebogI+zOZ1OVn37LaaKCiJkmVxFYWdGBmOmTEGn6765IA/ecQd/dzj239WPUFXinU4e+MMfmLJtW7e9Tm9j0eq4IPMUSh3NNPs8nGYwEW3s2MZ7R3JGciZ35m9jgeInCnADd0gyp8UkHfVLbG1lEcPsTZxvMLV86asqn9sb+bmymNMS2455ljpsvLVjPdttLV3TwyOiuX5ATre8h45IGTYOr3MxtqotICWBuou4/lasaV3b++1Y/PLzaua9/RLDa8vxl1exwAUDkvqS3cWn4dZKHTZ+ri5lr6Ls3413vKpwltvJ8qpizuziNivHS488z5rNZszmlj9Ep9PJzq1bqcnPR282k5qdTZ8+nRvktVqtfDR/PldfcgnRHg8qUGcw8Mknnxy1Qq2sqIhUjWZ/cgKICAkhqrGRqqoqEhMPLCB57x13sOntt9ngaNmZdEV5OZdffz2RVitnntn53UiPxYAB/UhP70Ntbe3+LeiPZ3Jyu90UFhRgq67GEhVFakYGJpOJrT//THJNDf33bTs/ENi0Zw87ExIYnH34nU07a9WmTXzRrm0mcMn27SiKckJVaXaWJEkkWw5sm5JvayC3qhSH20lkaDjZMUmd/sI/J7kv1Q4b6eWFDJFl8lSFrPBobssafsTzVFWluK6S6/XG/X+PkiQxRW/gsdpKaJWgnD4vf1m/jPt8Hm6mZSztmYYaHt2wjOfHTkUbgP9TrcFEv4nn4myswutsxhg2GYOlY3tVHSt7XTnOwq1IXheOsFQUZSC2pkZWv/ZPnooOI0Z209RkY7zfz7Mle0gJCSeqm6oCdzbVM1mS9icnaNkbYZbi58v6apGgDsXlcrHiiy9IaWhgTFQUzsZGcr/8EvvppzNw6NBOXeuMM86goKKCNWvWIEkSo0ePbvP0cziqonCor3dJVdsMCtvtdt586y22O53798Y9DXjK4eD5v/wlYAkKWqrwWifO48Vut7Pyyy+Ja2gg3mSiYft2lq9fz6jzzqNqxw6Gt9uhNzMmhlXbtnVrgkqyWsmrrKT1fe1OIC48/KROTu1tq68mvyiPCzUa4mQtufXVfN1Qy5n9hnYqScmSxLVZw7kgfSBF9iauMppJMIUc/URaxjjaP2PJSKi0HYdaXlXCWMXPHfv+bQD+oqp87/OyvraCMTHH/28bDlThEYAqvLo9Gwhbv4CpkoRRllnu3EC+vpY89UyG+33EmkyotpZjw2UNY1HJtzV0W4KK0htZjIRK2017dkgyER3c8r4n9OgnfO+ePcTX1zMgKYkQk4mY8HDGJCZSuGbNMRUqaLVaJkyYwPjx4zuUnADiU1Io8vnw+Q8s8NjsdFK7r/vxV1VVVYTJ8kEbtw8HCvLzOx1rb5C7aRPpNhtDk5NJsloZnJjIAJ+P7fuKFNondkmSoJsHxW+96y5uNZsp3ffvCuB3ZjO33nHHkU47qSiqypayAn6j0zNYZyBao+FUg5FzVYVfqkuPfoFDiNAbGRoZ2+HkJEkSyZExrPC427Sv8LhJaleeXuloZvQhFlQdqShUuo68dXtv5HM7kTd+z43mMEaERjHIEsHlIVaSczdSUrDn0DfI0K1Vk6dExmDT6XkK8NIypvgN8B9JYnJS8E456dEE1VBWRmy7ZY70Oh1hikJTU2DWhouOjiZ23DiWlpWRW1LC1pISfmpsZMjZZ7fZTTcpKQmnLJPb7vyFkkT2iBEBiTXQqnftIiW67SZ2yVYr9QUFWPv3p6C6us3PdldXkzBoEN3pD3/8I2ffeiunmEwMCA1loNHIuBtu4N7Obqx1ArP7vBj8XhLaFRoM0Gqpbw7cGouj41NZZjTxhsvBUpeD111OVhjNjIpLaXNcemgk38qaNs9VCrBQlkgPCXyxz/FmrysjS1UJbbWhoSxJjNLqUB12Nmo01Lhc+39mUxTWINE3rPsm0MuSxP3DT2deaARxskyyrOFmvZE/DR1PbIDG/Y5Fj3bxmSMjaSoooPUDtqIoNCsKJlPgZmQPGT6cpLQ0KisqMGg0nJacvH+M7Fd6vZ4HHn6YCx96iBccDoYA84HHTSYW/vWvAYs1kLRGIx6vt834nMfnQ6PXc8ro0ayuraW2uJgISaJGVfGnpjJuyJAjXLHzZFnm0aef5u4HH6S4uJjk5OSAb1cf7IwaLQ5JxqYohLbq9iz3+zFbOvYE1B1CdHou7DeM3bZ68txOIg1mLgyNOGhMaUxMIp8XbOe3Lgd3qgpe4FFZxmgOY0g37eobTDRaPc2HeBpqVvxExMSS9pubefLdlxlZW4u72cEyj0TfpIxu6977VazRzCMjJ1PjcuBRFBJMlqBfxKBHE1Ra//6s2riRiMZGYsLD8fp8bK+oIGLQoIAvIBsZGXnUgorb77yT6Lg4HnzsMYrLyxk9ciRfP/kkBQUF/OOJJzCYTFx+/fWkpaaSv349jro6IpKTyRo1iujo4Png1dTU4HA4SElJOeIfaMrQoWxftIhRffogyzKKorCjooLkceOwWCxMuvBCysvLsdvt9A0PJy4u7rj9wYeEhDBw4MDjcu3eTifLZMQk8WFFEbMNBsJkDSU+H18ofobGBnbbB60sMyD8yFM7dLLMwzmT+Ch/G2dUlaKRJMbH9+H3SRl8sjeXclsDA+rCuTgqkgVfL2HLy5/Q4JWpSMlG8fcBguOOX1UVXI01aI0WdMbDf19ZrIkUmMPZ5Wiin7llyagGn4dlBombxkwgNj6R80dlsXHeG5QuXc1kR/cVRxxKoColu4PUmX7OkSNHquvWrevWAKqqqti2fDmeqioUrZb4U07hlJEjOzyG1JMURWHORReR+8MP3Gi345Ak/qHXc8WkSdxzzTWEmUxUNjSw1eNh5MUXExV1iF07j6CpqYknH32UT//3PzQaDZdecw1333//QU93HVVRUcEVV/yWlSt/RJaNxMRYeffdfzFp0qTDvr9Na9ZQs3kzEZJEo6oSPnAgIyZM6BX/P8eDJEnrVVXtUi1yZlik+tzIyd0VEgB+VWVtZTEF1aUYFAWfTk92YjoDeskTSanDxiPrlzLD72eSqrBckvhWknh4QAYXZudQ7Yb3dxcxj3D6nnMdktS50YmGsl2UL/sftqpCLOGxxJ42m5i+R65MPJLKvLVs/+6/+Dx+VNVOdEYOp0y/7rCJytlUg23FR6TYGzADW/0S6Vf/jnuuuxBVUcmJVFHWfMWeuV9RUHjMYfUaFyz5tEOfox7/lomNjSV21ixcLhdarbZXffEtXLiQX374gXV2O0ZaBqtT3G4+XrQI7yWXoAkJIdFqxVddze7Nmxl9xhkdvrbP5+PsCRPot2sXc91ufMBTzz3H+YsWsXDlyk4/qaiqyplnziQvbzI+31zASFHR18yYcSlbtqw55Np8siwzYtw47EOHYrPZsFgs3dq9pigKixYtYuvWrfTv359p06ad0CtiHE8aSWJcfB9GxSbj8vswa3XIQd5909oHuzbzfz4vd+/79yBVxaSqJJVXEzFKj1eVODs6nk07d9NYXUxobMcn1DaU5rHjg0d52ufhbGCd08Ztn/0N/zm/I37waZ2OtbF8D7/MfxvF9ykttbxN1OT/kU2fvcqoy/90yHNMYdEYz7mJuvoKanxefFI4cQO7b+y62ethdU0ZLr+f4VGxJJlPjG7woKnTNRqNvSo5ASz48kvm7EtOAB4gEhgry2zcuHH/cdbQUGyVlZ269vz585H27uU9t5thwEhgrstFxdatLF26tNOxrlmzhsLCRny+J2hZCEoCZuDxXMPLL79xxHMtFgvx8fHdmpwaGxs5dfhw7rroIvbeey9/vfxyRmRlUdnJ35PQllaWCdHpe1VyAlhfX80Nrf5dBpwBlNQ37i8MlZDoB5RUdG69vLLF7/OCz8PvgAzgUuBjn4fiJe8fU6VcwZqFKL57aUlOAGGo/pdpKM3D0XD4v19JkrFEJRIam4qsbSnAUhUVVfGjdmAZqcPZUFvBLSu/pWTnZpTdW3jo50V8uHvLMV8vmPSujBBkQsPDqdXpYN/agnpaqpFqZblNN1x9czMhqZ1bQmX9zz8zrbm5TQmqBjjb7WbdunWccZSnMa/XS2FBATV796Izm9mam4skDaZ9cbjXO5jduxd1Krbu8ODddzMwN5c3PJ6WklqPh7ucTu68+Wb+8+mnAY9H6FlmWUOt38evneBxwBIgTpZBVYgxyKgxFupqQ5GNYRTXOzp87drKvZzTru1UwO5oorCyFs1R5gF5bbX4izajs9fiikqhoaoSaF8MZAA5neKycszq0W/kxk8aAiqoip+ccB/en+ZT+MXCTnfvufw+Xty2hq8VPxP2tT2tQk5pPqdEJzC4l3TxHo5IUF0w59prmfDCC1zn9TKIlsfRcmCJonD74MGoqkpVQwPb3W5GdHLicWp6OgssFrC3nRfyi9HI2KMkO5/Px0/ffktYSQkZoaE4PR5Cy8rwuBcDzdBqPrnZ/BVnnDGxU7F1h7kffMDqfckJWtLmAz4fifPn867fL7r6TjKTElK5u7SAuaqCHhgK3A5MM+ppLstHkjWsUPWYswcx5exzkDQd7/xp+jKezeWFtF44aA+g1xuYcFbOETcYrC3chW3eZ5ytqsTqjeyu2cwnJie5mrmo/qmtjixDUndz2nlnYgg5+t5Z14zu0+XkBLC5ropspP3JCSAa+L3iZ3lFoUhQJ7P+/fvz93//mwm/+x1jdDrswC5J4oWXX2azJOEqLSU8KYnsqVM7XcV32WWX8Zd77+Vlh4MbVRUFeFGWKbBYmDlz5hHPLdy7l9DiYka0WjIqJjycGTnZfLPpLJzORwErOt0bWK1bue66tzv/5rvIryi0X7FPR8s43omyrYPQcZdmDOFFu40+jbWMliTWqSqJIeFUWhP4Q14B5oRYho08hUfvuY+dXgtSJ3atTbzrXm594E986nRyClAIXG0ycel1N3Ht2MNPUlVVlfc+f4E74qPoF9ayDNLZQJ9CmXvqPqXRE4vivwIoxGi6n0uuvYnfTu7YNIvuSE7QUhyjP0S7gZYx3t5OJKgumnP11Zx/wQUsXrwYo9HIGWecgcHQshKxqqrHXHYdGhrKwhUruPmqq7hvyxZUYNzIkSx8//391z+c2qIi+rQbLzIbDNw+51JGzbTz7rsP0txs44ILzuHBB5f3yLyiC2bO5IV58/ib78CuRi/KMtMnT+51Y5FC1xk0Gu4adipF9iaK7U1MMYeSFtKSFFRVJT0VJKeTkB3LGDlmBtDxm5ica2cR5aln6pNPobg9eGWJ395wAw88eDcazeGv09DYSEhdGeMSEpCkA8dNS47jl9goGiyVLF56KdaoKO649WYuu+TiNscdierrenICyI6K5RVVYSsHOh2bgZdlDZe3myDdG/V4mblwdLW1tciy3OGt2TetWUPYli1ktForT1VVFpeVMeqKKwgL6/ntuysrK5k0ejQpdXVMam5mTUgIv5jNLFmzhrS0tJ4O77CCtcz8ZJC+r2e77+wZx3S+z++nptFGZIgFg/7oK+67PF4eeuJlnggPxdyqy3mbzc78WCt/uvHYdzHwudxdTk6/+rGiiLfyNnC5ClGqwn9kDYNikvjtwJygnYjba8rMhaOzdmBPq9b69OvH+o0biXU6CTGZUFWVXRUVGFJTgyI5QcveYBvz8vj444/ZunkzMwcO5H+zZwd8grbQe+z/Mp/7VZeu05nV/hKA97bu5BKLCZ0kYfMrfOB0MTjUzJ4uxtFd851Oj+9DVriV5ZXF1Ph93BydQFZYVNAmp84QCeoEFBUVRdY55/DT0qWY6+pwqSrmjAxGTgx8McSRGI1G5syZA3Pa7+EqCIcXyImsWYY0lhn93FlbTawkUY5EZnwaYU5rUE2ojTNZuDhtQE+H0e1EgjpB9UlNJWnOHBobG9Hr9YSEBG5NNkE4Ueg1Gs5K7U9TYirNXi9jDcaA7v57shO/6ROYRqPp9PJKgiAcLExnIEzX/duwC0cWNCtJCD3P7/dTV1dHc3NzT4ciCL1ak9dNuaMZt9939IOFwxJPUAIARYWFbF+yBIvLhQswp6eTM3EiRqPxqOcKgtDC4/ezrGQP9Q37xqwkiX5xfciJSTwhihYCTSQogbq6OvK++YZTrVZCrNaWqr/CQtYvW8aEqVOPfgFBEABYWb6XPvVV3G5sqfprUhReLytgp8FEVrjobu8s0cUnUJiXR6ZOR8i+TSIlSaJffDzO/HxsNlsPRycIvYPH76esrpLzjUZ0+56WwmSZGVotu6pLezi63kk8QXUTp9PJ4sWL8fv9TJ48uVdVzXnsdkz6tgumSJKESZbxeI59lWVBOBaVTju7bfXEGMz0C4vsNV1jHsWPQVUxtFuQOUKW8XRhtfKTmUhQ3WDhwoVcOWsWgyUJLXCtz8dr77zDxZdc0tOhdYi1Tx/Kdu4kvtVKFQ63G5tOR3h4eA9GJpxM/KrKGzvWs7q6hAmSzCeoGIwW7hl2KhH64B8LtWh1qAYjBX4fGdoDK1Vs9nqJiYnpwch6L9HF10X19fVcfuGFfGKzsaSpiYVNTSx2OLjpmmsoKirq6fA6JDU9naakJDYUFVHV0EBhVRWrqqvJmjhRrIsnBMx3pfnUV5dSqCjM9/vY7fczw27j1W0/93RoHSJJEjnJfXnb72e528Uur4cvXU6W6vRkxyT2dHi9kvj26aLPPvuMyRzYugxgGDDb7+fDDz7g7nvu6aHIOk6n0zHh3HPZm59PfmEhOouFof37i7s+IaCWl+7hRcXPr0sXS8AjqMQ11mLzegjVHWrd7uCSHhKBpf8wNtaU4/C4iLSEMT0qjpBeEHswEgmqi2w2G1bfwXMdrF4vTY2NPRDRsdHpdPTLyqJfVlZPhyKcpJx+H+1XnTQBBqllY77ekKAAYo1mYpP79nQYJwTRxddFU6dO5TNZpq5Vmx34wGzmnOnTeyosQeh1sqMTeb1dQcRXtKziEG0w9UxQQo8SCaqLBgwYwLU33cQYi4W/A/8ExlosTJo5k/Hjx/d0eILQa1yYNoCv9EYukjW8C9wpSVwja7h+4MheU8kndC/RxdcNnnz+ec6aMYO577yD3+fjySuvZPr06eJDJQidEKE38vToM1lcXsj7DdVEmUJ4MimDeJPYguVkJRJUN5AkiSlTpjBlypSeDkUQejWzVseMlExIyezpUIQgILr4BEEQhKAkEpQgCIIQlESCEgRBEIKSSFCCIAhCUBIJShAEQQhKIkEJgiAIQUkkqG5gs9moqqpCVdWeDkUQei1VVWnwuHAeYukw4eQk5kF1QX19Pbdcey3zFyxAJ0mkJCXxjzffZNKkST0dmiD0KpvrqngndwP1HhdeYGx0PNcPyMHSatsK4eQjnqC64JJzzyViwQJKPB5q3G4ezc/nkunT2bVrV0+HJgi9RrG9iX9sWcVLbgd1qkKJqpBaU8E/tqzq6dCEI3D4vNi8x3cjRvEEdYy2bNnCzl9+4TuPB82+tpnAGo+HV198kb+99FJPhicIvcZ3xbu5TVX4dWnlCOAVVSG5qZ5Sh40kc+iRTj8prawq5T/526h2O4jSm7g4NYuzEtMAWFFZwvv523D4vJyVmMacjMHIHVx27XDn1rtdPLx5BQCP7ttA8vuyvUyMSzlebxEQT1DHbO/evQzRavcnp18N8/nYm5fXIzEJQm9U67AxrN34rR4YIElUuxw9E1QQq3Ta+WfeBm7sl81rY6cxJ2MQr+7cyO6menY01PL8jp85PyWTZ3ImkddYxzelezp03SOdu6qmlIyQCDJCIlhdXYZXUbD7PEQaju9OxyJBHaNhw4ax2u2mqV37t0YjOaef3iMxCUJvlBoezddS26+iOmCzopBqCe+ZoIJYfnMDiSYLw61xRBqMnBaXQrTRTInDxvyS3WRHxjA9uS8J5hCuyRzC1yUdS1BHOtfm9bTsc2U00+T1sLyymAmxycfzbQIiQR2zlJQULr/iCs41m1kG5AL3aTQsCQ3ltzff3NPhCUKvMS2lL19qtDwA7ASWANNkDZMTUo/7HXpvlGYJp8jexLqactx+H8sqimjwuBgSEc0eWwPDo+L3H9s3NJJql4Mmr/uo1z3SuRatDpvPg83nWZiLtAAAIABJREFUwazVUmRvIi3k+N88iDGoLnjpjTf4V3Y2d7z4Io02G2efey4rHnsMq7X9vqCCIBxOhN7IY6Mm82n+Nt6pqyRUq2NSciZTkzJ6OrSglGAOYVafLB7bV0QiAX8aNIpooxmH39tmexKNJGHS6Kh1OwnTGY543SOdm2ON56O9y5AkuCJ9EEMjY47Le2tPJKgukGWZ226/ndtuv72nQxGEXi3WaOZ3g0b1dBi9Qr6tgQVlBTw+/HT6hUaytqacl3duJMpgQiNJ6OW2HWN6jQaXz3/U6x7p3PSQCF4fNw2A/xZsJ1Sn57erFjAprg9XZAzqvjfXjujiEwRB6EWWVhZxWmwygyOi0Ws0nBqXTE5UHIvK9xKmM1Dvadud5/R50clH/6o/2rl6jYZKl50USyifFu7k8vRBzC/Zjdt//CZWiwQlCILQiyiqSr3H1aatweNGUSErLIodjTX720sdNhx+H1aD6ajX7ci5yyqLmRiXQpPXw8Bwa8vY1HGcCyUSlCAIQi8yJCKG1TVlfFKYx4rKEl7J28iWhmrGxSRyelwKP1YWk29rQFFV5u7NJTM0Yn+xidPnw+M/dHff0c5t9Lgxa3ToZA0hOh2FzY3YfV4sWv1xe69iDEoQBKEXGRuTyFUZQ1hQmk+Vy0GYTs9vMocyJiYRgGmJGdy1fgkhWj2KqvJQ9vj95z62ZSWZoZFcl3nKQdc9JTLmiOeuriljSkIqAOclZ/KP3PVMSUjFpD1+aUTqzAKnI0eOVNetW3fcghGEYCdJ0npVVUd25RqZYZHqcyMnd1dIgnCQEruNUoeNgeFWwvRHrt7rznM76oIln3bocySeoARBEE4wyZZQki3HtkRUV87tbmIMShAEQQhKIkEJgiAIQUkkKEEQBCEoiQQlCIIgBCWRoARBEISgJBKUIAiCEJREghIEQRCCkkhQgiAIQlDq1EoSkiRVA4XHLxxBCHqpqqp2aTMc8TkShI59jjqVoARBEAQhUEQXnyAIghCURIISBEEQgpJIUIIgCEJQEglKEARBCEoiQQmCIAhBSSQoQRAEISiJBCUIgiAEJZGgBEEQhKAkEpQgCIIQlESCEgRBEIKSSFCCIAhCUBIJShAEQQhKIkEJgiAIQUkkKEEQBCEoiQQlCIIgBCWRoARBEISgpO3MwdHR0WpaWtpxCkUQgt/69etrurqjbpjeoMYazd0VktANDHowRIWDJQKHD2rtHpptTvTaY7+H9/gUUhPCAbDoNOC04W1qwtHkBsAcZkAXFgamUOxeP/UOLy67q1veT7Brqsjv0OeoUwkqLS2NdevWHXtUgtDLSZLU5a3aY41mnhs5uTvCEbpJeir0nT0DecwM1tdLvLu2iJVLt5ISeew3EsX1Dl67dzoAOQkhaHb8SPm3C9m4cA8Aw8/qS8I5Z+EfeDrry5v5eEMpO37e2S3vJ9h999QlHfociS4+QRAEISiJBCUIgiAEJZGgBEEQhKAkEpQgCIIQlESCEgRBEIJSp6r4uqK2tpZdmzZhq6jAHBVF5ogRxMXFBerlBeGE4FdVfqmtoKCmHJ/iJyEimuExSYTo9D0dmiB0u4A8QdXU1LD+k09ILiriNKORjKoqtnz6KWWlpYF4eUE4YfxYmo+nZDc3+L38EehbXcY3e7bg9vt6OjRB6HYBSVC7NmxgsF5PcnQ0Rr2ehKgohoeFkbd6dSBeXhBOCA0eF7W1FVxnNJOi1RGt0XCO0cRgt4vchtqeDk8Qul1AElRjaSkxYWFt2qxhYTiqq/H7/YEIQRB6vRqXk0wJdJLUpn2ALFHvsPVQVIJw/AQkQZmtVhodjjZtNocDfWgosizqNAShI8L0BkpUUFW1TXuJomI2mHooKkE4fgKSHfrm5LCloYGmfUnK7nKxqbqajNGjkdrdDQqCcGgxBhOERvCF24lTUVBUlU0eN6s0WgZFdml5QEEISgGp4ktKSsI3fTprVq/GX1qKZDbT96yz6NuvXyBeXhBOCJIkcWZqFqvK9vJwfRWSqhJqCeOMpAxRxSeckAJWZp6ank6ftDS8Xi86nU48OQnCMTBqtJyRkokvKQO/qmDQBOwjLAgBF9C/bkmS0OtPjjs9RVGQJEkkYuG40Moy2pNgnr2qqqiALD5HJ6UT/y88wBYtWsTYwYPRajQkRkby+COPiEpFQegkn6LwwZ6tXL98Phcv/YwH1y5ie0NNT4clBJhIUN1o7dq1XH7eedy9fTteYHFjI98/+yz3/+lPAY/lyy+/5LTTziUzM4ebb76DUjEpWuhF3s7bQE3JHtb4fbiA++yNPLv5JwqbGwMah9dlZ9ePH7P8tQdZ+dZjFG9ahKoqAY3hZCYSVDd67tFHedjl4iJAAwwE5jocvPbaa9hsgZun8txz/+CKK+5kxYqr2LPnZd54Q0t29ljKysoCFoMgHKtGj5vlVSV8rPjpB+iAy4G7FD/fFOYFLA7F52H1e4+zdw046l7DVvUYeYt+Zus3bwcshpOdSFDdKG/bNsa1m6MSD8TpdBQXFwckBrvdzkMPPYrdvoCWj/UYfL6/0dR0Ic8880JAYhCErqh02UmTZCLatY8Hyu1NAYujafdq3LYUFP8HwGnADPzepVTsWIu9rjxgcZzMRII6Rm63m8LCQpxO5/62wdnZ/NhuMLcUqPL56NOnT0Di2rFjB1ptH6Bvm3av9yIWLVoVkBgEoTOavG5q3c79E5DjTRYKFIX2izctRSIppH3aOn6aSwvwey8BWn+mQ5CkyTSWnRxbs/c0kaA6SVVVnnn8cZKjozl10CCSo6N5+N57URSF/3v4YR43mXgfsAHrgFlmM7fedhshISEBiS8uLg6PpxhwtvtJHikpCQGJQRA6os7t5KmNP/L7n77lrtXf8X+rv2NHQy1hOgOTE1K5QNawCWgC3gT+IctMT80KWHy60DAkzY52rSpIOzGERAUsjpOZSFCd9Oorr/C/J55gVXMzxQ4HGxwOFr70Es8+8QTDhw/n8++/5+2RI4nVarksLo7LHn2UR558MmDxpaSkcNppp6HX3wHY97Vuxmx+jLvv/n3A4hCEI1FVlac2LueMhhqqVIVqReFpl4NnNq+g2uXg6v7D6JuaxXSdnjhJ4tWwKB4YfjrJltCAxRg58FRk+X1gAaACXpCeQWdsJCp1cMDiOJmJWX6H4fF48Hq9WCyWNu3/ePJJ3nQ4yNz371TgdYeDs557jrsfeIAJEyaw+OefO/w669atY8mSJVitVmbNmkV4eHiXY583722uvPK3LFqUjE4Xg0Zj44UXnmbSpEldvrYgdJbT50Mny2hbrbu5vbEW2ePicQ50oM0CFqsqi8sKmJ0xmAvTBnBh2oAOvYbb72dNTRk1Lgf9wqIYEhHd5TmI+tAYhs+6jS3zb8TnUVFVByHWFIZddA+SJO7tA0EkqHYaGhr440038dHnn+NXFEYOHswLb75JTk4OAEWVlQxpd85AoLKxEb/fj1bbsV+poij85oorWDJ/Phd6vazW67n3jjv44rvvGDduXJfeQ0REBF9/PY+qqipqamrIzMzEZrNx11338/nn3xMeHsbtt1/HnDlzxERi4bjZUl/N+3kbKXLZ0SBxRnwKV/UbhkGjocblYAhtR3cAhqkKHzuaO/U6pQ4bf92wjMGKwlC/n/c1MqEhEdyTfSp6jaZL78GadgoTb30OR105ss6AKSya+uIdbPv2vzjqqwhP6EPfU6cTEp3SpdcRDk3cBrSiqiqzpk1D//nnFHo8NPl83LB5M+dMmkRJSQkAo045ha/bnbcAyO7bt8PJCeDDDz9k21dfsd3h4O9eL5/Y7bxhs3HFBRd028Te2NhYBg0ahMvlYsSIU3nxxRp2736R9etv5+ab/87tt9/TLa8jCO0VNjfy919W8pSzGbuqsktV0FUU8+/tawHIDItkqargaHfe57KGjIjOLXz76ra13Of18IPfx/OobPf7ibfV82VR9xQySJKMxZqEKSyayp0/s27uS9TuvQpn4wdU5E1h9bt/pamyoFteS2hLJKhWNm7cSMHWrbzi8WClZf7FNcBlHg+vv/IKAH95/nluN5t5FcgF3gF+Yzbz1xc6V8I97403uMNux9yq7XzA4nSyfv36bng3B7zxxltUV5+Cx/MaLcW6M7Hbf+D1198Qc6OE42JB0U7+qChcSMuXTALwH1VhU10l1S4HSeZQhlsTmCprWAJsBG6SJLbp9EyM73jFa4PHxV67jZtbtWmA+xSFNRWF3fmWUFWV3IVzUXwfADcBOaDeh9/7KDuXfN6tryW0EAmqlT179jBclg/6pYz0eMjfvh2AiRMn8uWiRXx31lmcHx/PJxMn8uE33zB9+vROvZaiKByq80Gz72fd6YcfVuF0XtCuNQq9fhzr1q3r1tcSBIAqh43RtJ0TaAb6STJVrpbnpt8NGsWAjMH83hTCRQYj1Ynp/HXkZEyd6IlQ1ZYvsfZdhVqgu9d78LkduO3VwOR2P7mQhrLATSA+mfTaMSi/38/GjRtRVZURI0ag6WJfM8DQoUO5zdeytIqxVfsPJhNDx47d/++xY8fy6fffd+m1Lrr2Wl5at44L7HYM+9oWAvV6PSNHjuzStdtLS0tEo8mjbc+hgt+/k8TExG59LaH3sXk9FNmbiDaYiDNZjn5CB/QJi+L75kbObjVxvRbIVRVuMbdU4mlkmRkpmcxIyTzMVY4u0mAkyRzC2/YmbtzXpgDPyDKj47p3XEijMyDJGlSlBGh97Tz0ZlF2fjz0ygS1atUq5lx0EQa7HRloNhp57+OPOf3007t03aysLKZMncqF333H404nVuBNWWaZxcILN9541PM7Y86cOXzz8cdkL13KxXY7RSYT38gyH330UafGsjri1ltv5N13J+FwTAYmAm602sdIT7fuL/4QTj6qqvLhnq18U7KHLFkmX1UYFBHNLYPHYNLqunTtc/r0576KImL9Pq4ECoE/yRqmxKcSrjcc7fRO+e2gUTyw8Ue+QWGoz89XWi1yWAgvnZqFuYNvI3XmWUg5Z7OuVuG9n0tZsXQrcrvHMlmjJTl7CiWbb0Dx/Q+wAnuRdX8gY9yZ3fqehBa9LkE1NjYyc+pUXrfZOJ+WR/sFNhuzpk8nd+9erFZrl67/9ty5PPPEE8x+9VVsDgfnTJvG8mefJSqqe++QtFotc+fP58cff2TxokWMionhucsuIyam+3dGHTRoEPPmvc11112LwwF+fxMjR45m3rzPRBXfSWxR+V5yS/PZpSrE+RVcwI31NbyVu4Fbhozp0rVjjWYeyZnEvN1beLqxlnCtjsnJfZme0v2blKaGhPPF7Gn86HBTVt/ErX0SmdQ/HY3csREM6/CBSDlns75R2yY5pUSaDzo2a/Kl+L3/pXxbOpImFlWpIX3MDJKGntHdb0ugFyaoTz75hNMVhZmt2qYBU/1+PvzwQ2655ZYuXV+v1/PnRx7hz4880qXrdIQkSUycOJGJEyce99eaPn065eXT2L17N6GhoaJrT2BR8S5eUPzE7fu3EXhRVUipKed6n69TY0GHkmIJ487sCV2O82jSU6Hv7AsZeozndzQ5AcgaHUPOvZasyZfgbq7HFB6DRte9T4TCAb0uQdXU1NDH7T6oPdXppKa6ugci6j00Gg1ZWYFbKkYIbk1eL+3r5SIAgwROv7fLCSoQWpLTDKScs5G0x7YZavtuvcMlp9Z0Rgs6Y/eM1wmHF/x/ge1MmjSJywwGHvf5MO1r8wCfWiy8cpKslOB2u/niiy8oKCggJyeHyZMnIx+hO6O+vp6NGzcSHx/PoEGDAhipEMwGR8bwv6oSHm3VtggI0emJ1BsPd1rQaJ2c1jdqkdoPGnVARWkxz73xP/J3lxOelkNqnyMXVjTXFONubiA0Lg29KXDLLp2sel2CGjVqFBOmTuWM777jDrsdDfCixcKQAHWV9bSCggLGjz+T5uZ0XK7hGI13MmBAJEuXfn3QskwAjz76FE8++TQGw1C83nwGDuzL11/PJS4u7hBXF04mszIG8+e6Shr8PqarKpuBZ2QNN/cfHvRjk+2T03s/lx5ca34U27/7mDX//TeKMhPQUbP2L2jOuJTUnLMPOtZtb2TDRy/RXFOFrElD8W0lfdx5ZJ56Yfe8IeGQel2CkiSJd+bN47///S//ffNNVFXlumuvZdasWXg8HgyGE7s/+Oqrf09V1U0oyt0ANDc/zZYtV/Doo0/y9NOPtTn2888/55ln3sHl2orLlQT42bz5AWbNuoYVKxb0QPRCMIkzWXhq9Jl8U7yLBxpqsJpCuCupL3EmC35F6XCRQaClpx6oums/dtRRnqZq9sx7GdW/EUgHQOUhdi4ZQUzfbMwRbW/gNn/+KraqM1GVp1F8GqCMvWsmEhqTSFxW1wpKhMPrdQkKWsZSrr76aq6++mrKy8vZtmwZ6z74AK8kYR04kGHjxqHTda1MNhg1NTWxZs1yFOWzVq0ybvc9vP/+ZQclqOeffwO7/SEgaV+LBp/vUdavT6KoqChge1QJwctqMHFV5lB8isLK8r1sKMojDJVmjY4hCakMiQquJ+1fk5PWaEDS6lu69SQ6PHb0q707fwH1Un5NTvuujqpeTFXeWtLGnLe/1dVUS2PZblRlOeyfXp+I3/tX9q59XiSo46hXJqhfNTQ0sGX+fEaFhhKZlITf72fHtm2s9/kYO7n9bO/e78AKE+3vbDWo6sHz5mtr6ziQnH6lR6eLob6+XiQoYb9VFYWEV5dxg8GISZap9vt4vXgXZp2BjNDAbRLYEVqjAX18Mr4uXENF4ZBff6oOFVebJq+rGUkTBf72vTPJeJydW9hW6JzgfIbvoMKdO+krSUTu2wxQo9EwKDER286d2O32o5zd+0RERDB8+Bgk6d+tWlUMhue47LKLDjp+xozJ6PX/bde6HlmuZ+DAgcc1VqH38Cp+imoqmLUvOQHEaLTM1GjIrS7t4eiOj9h+o5HkD4GSVq0lSPKHxPYf3eZYizUJSWqiZQvSAyT5XWIzxefoeOrVCcrZ0EBIuzEnWZYxSxIul+swZ/Vu7733Mlbr37BYZgIPExIynszMXB555P6Djr3rrjuIiVmK0Xg18AWS9Bxm83n861/Po9cfW0mucOJx+f2YUbC0G3OKkTW4PCfm58gSlUDfU89H1mYjyX9Akv+ArM2m76nnY4lsu/O0rNEyaOqVyNpzgWeBL5C1c9CbvyV97Lk9Ev/Jold38UWmpFCxZw9xkZH721weDzatlrCwsB6M7PjJyspi794dzJs3j/z8AnJy7mHGjBmHXB4pOjqaLVvW8K9//ZsFC14jJSWeP/7xc0aPHn2IKwdedXU1TU1NpKenH7FMXji+LFodXq2eEp+P5FZ/R9u8XiK7uDJLMMsYO4PYfsOpzFsDQHzWQ1is7bvEWyQMmoA5Mo69Py/E1fQx0el9SRnx16AoNVdVBWdDFVqDCb256xueBpNenaDS+/Zl+datbCstJSk8HKfHQ15zMxmTJ5+QRRK/slgsXHfddR06NjIykj//+T7+/OfjHFQn1NTUcMUVN/Ljj0vQaEIJCdHx5psvMmPGjJ4O7aQkSxLZiRm8WZjL+YqPWFlLrt/HDxotU2MO/YV9ogixJhEy/uDu8UMJT8gk+/xjX9j2eKjes4Gt37yLz62CaicieRDZM284YRJVr75t1ev1TDjvPBg3js0GA3sTEuh/4YX0F+MrQe3ccy9l6dI03O4yHI4iqqreZvbs37Bly5aeDu2klRVhJSfzFH4IjeJNrY7t1nim9c8m0hD8E3ZPVs3VxWz67FU89v+g+EpR/BXUF49n3dx/9HRo3aZXP0EBGI1GBmdnQ3b2YY9xOp28/dZbfDtvHiFhYVxzyy1MmzYtgFEKv9q6dSvbtu3G613IgZLdibjdf+CFF/7Nm2/+qyfDO6mlWMJIsRy5a7zUYeO7ol1U2ptICovknJR+xBg7Xt4tdJ/C9YtQ/bcCvy5Ua0JVnsFel0ZT5V7C4tJ6MLruEdQJqqKigj0bN+KoqyMiOZl+2dlERHSu5NXtdjP11FOx5ObyG8f/s3fe8VFV6R9+7p0+k957IY0eqvQiXRAbqNjL2gs/2666uru69rWuXcTeVkUUFZCi0juE0BIgCaT3Nsn0uff+/hhaQktIAgHm+XzyR07mnPvezMx9z3nPe96vlRpg5p9/csODD/KPZ57pGMO9HJfi4mI0mjRoJtcoSd3Izd1wZow6x3FKEhnVJRRVVwAQExxG3+AotK3UUMuur+alrau4R5a4HfijoZbHSvN5qv/okzo2L+2Pta4ORenZrFWFIKTgaKyBc8BBddoQX0F+Pjt/+IGkykqG63SE5eSwfs4c6urqWjXOt99+i2r3buZbrUwH7gBWWiy89vLLlJaWtpu9ZWVlLFq0iOzs7HYb81ykb9++OBybgKaFffX6nxk3bvCxO3k5ZWRFYdH+LHxL8rlbkbhbkfAtyWfR/ixkRTn5AEfw1e4M3pElngUuBl5TFP4uufk2Z1u72asoCrkNtWytKcfidrXbuOciQXEJiOq5zVqrUaTN+IUnHrPP2UandFCKorBn7VoGBAcTERiIQacjPiyMroLA3m2t+zL88csvXHtA2PAg4cAorZZVq1a1i61/vf9+uiUk8PLVVzO2Xz8mjRjRakd6vhAWFsY999yNyTQB+AXYhEbzfwQErOTuu+840+adcxRYzJga65ihNxCpUhOpUjNDb8DUWE+hxdzicVyyRJbFzJXN2q8HMuuq28XWcpuFR9cv4a2MFSzesZ67V89nfuHedhn7XCSu3zg0+pUI4r14zmj9iqgZQ0yfMeh8Ak/W/aygUzoop9OJVFeHf7Pip6H+/tQXFR2n17EJCg+n+BihjCJoFxHCD2fNYtnHH5PjcLC0vp79NhuJGzZwz003tXnsc5VXXnmOd999kL59XyU+/lbuvFNNRsbqNotNejmaKruV7gpNir8KgkA3RaHSYWvxOCpBRC+KlDVrLwL820GWQ1EUXtu2mr/YGsmTJFZLbrbKMgvzdrG91iujcyw0eh+G3PIUsX2L0fvPwDfsCbpPGEHXcdeeadPajU65B6XRaFD0emwOB4YjDuKarVaMYWGtGuuWO+9k3Mcfc4XNRm9AAT4SBOp8fBg9ejT79+9n8+bNxMbGMnDgwFZXcf7ojTd43mrl4KNVA7zkdBLz22/U19fj739upHu2J4IgHKql6KVj8dXoKDzGZ7pAEAhshX6SKAiMjYhnZmk+XykyBsCMR8Z9THQXZEUhq76aBpeDrv7BBLRSrmO/pR6b3cYjHC5Kngg8Kkv8VJRLr8D2V5o+F9CZAug2/nq6jT/TlnQMnXIFJYoiCQMGkFFWhvWAOGFdYyM7Gxro0qdPq8bq1asXr8+axRiTiSF+fnT18eGN+HjmLVnCzNtvZ0C3bnxx661cP3YsQ9PTKS8vb9X4tXV1NNem9QX0okhjo7dOl5czS5JvAPt0elY67LgUBZeisMJuJ1+nb3WNvWuTe1MbFEaMKDJcpSFOFDGFRXNBaDQPrf2Nr7atISNrM/ev+Y3v83a2auxGl4twQTjqgRQNNLrOzWoWXk5Op1xBAaT16MFuYMWmTVBZiTowkLSpU4mIiGj1WNdefz2XT5vGhg0b8PHxoV+/fnzw3ntkfPsteXY7fnY7MvB4VhZ3XHcd85YubfHYYydN4vMvv+Ql9+HSlYuB4JCQMyar7na7WbRoESUlJQwePJhevXqdETu8nHnUosiELj1YU7yPX801AAT5BTExugvqVlbv0KlUPNB7KBV2K6XWRv5i8iVQq+dv65fwoMPG/XhWP2XA8MIcEv2CGBASeZJRPST7BrJbUdgDpB5oU4BPRZEeIWfusLCjsZaqvK0IKjWhSf28KrqnmU7roARBoGvPnqR2747b7Uaj0bRJRM1gMDQRNPz07bd5xmrlYHKsCPzL7SZy5Uqqq6tbvB/yxDPPMOzXX6lpaGCKw8F2lYq3dDq+mD37jIi+5ebmMnLkJBoawpCkNBTlKSZPHsu3336CqpVpxV7ODQK0eiYldsMhSYDH0bSFML2RsANnn/Y31uNw2LiPw6G5COBxWeLz4rwWOyiDWs31Sb0Ynbudv8oSscDnoordOgNPR52ZjLT8TYvZ8+e3COJ4EKwo8qf0ueweQpP7nxF7zkc6rYM6iCiKHVLYtKGhgZBmbQbAoFJhsVha7KBiYmLYtGsX7731Fh8vX05cSgp/PPAAPXs2P59wepg+/WbKyu5Blh880GJj4cKJfPDBLO655+4zYpOXzkFbHdOxsLndBB0jNBcC2NzOVo01IaYLsT7+/FacQ6PTQbeQSK6NTMTQDkkYraWxspA9y+YiS5kgJRxoXc/WnyYw+r43vSup00Snd1AdxcRLL+WjWbN423X4rMVCIDg0lNjY2FaNFRoayj///e92trD1FBYWkp2djSzff0SrAav173zwwXNeB+Wl3UnyDSAfyAQO1nJRgA9FFb3DYlo9XreAYLoFnPlszpKda5ClW4GEI1oHIYgjqczZRFTPUcfp6aU96ZRJEqeDx/75T5aEh3O1wcAXwGMaDTebTLz1ySdnJDTXHjgcDkRRT/MqDWDCZvNuNHtpf7QqFbek9mGcqOIpBD4Fxokq8gwmxp+h0Fx7ILtdoPgc1a4oPkjeA8SnjfN2BRUWFsaGHTv45KOP+G3ZMmKTk1l7770kJSWdadNOmaSkJIKDfbFafwYuPdCqoNe/wzXXXHImTfNyDjMyIo44H39+L8plpctBt+AI7gyP65CQ4ukiPK0/RZkfIbkexJOXC1AI8kJCk146k6adV5y3DgrA39+fBx56CB566Eyb0i4IgsDXX8/ioouuwOVaiMORho/PPOLjbTzyyIdn2jwv5zAJPv78pWu/M21GuxEQ05WIbj0oy+qN5PoLCFZE1WySR1yB3vfMhyDPF85rB9WRKIrCunXrWLhgAT59c2SEAAAgAElEQVS+vsyYMYO4uLgOv+7w4cPZsyeTTz75jPz8fYwadQfTp0/3Kuh6OWtxSG5WVRRTam1ggOTP7U4XHf1pFgSBHhfdTFTPXZRnb0RUq4js8bdzokL42YTXQXUAiqJwz6238tt333GtzcY+jYa+Tz3Fe598wlVXX93h14+MjOTvf3+sw6/jxUtHU26z8PTmZfSS3AyTJX4rUfH57idY/PU7BMV37LUFQSAorgdBcT069kJejst5myTRkSxatIjl33/PdquV5xSF95xO/rDZuOvWW2loaDjT5p1RNm/ezFVX3Ux6+kjuvvsB9u/ff6ZN8tKJ+Xx3Bne7HPwmS/wD+NMtMa6qlqdeO79D1rIsUbj1d9Z9/iLrPn+RwowlyLJ0ps1qd7wOqgOY+9VX3GmxcGQOUDowSK1myZIlZ8qsM86CBQsYOXIyc+b0Ztu2p5k920B6+mB27959pk3z0gmRZJkNtZU80Kz9YUnipz/brkRwtqIoChk/vE3275nUlzxBfcmTZP+xk4zv30RppYRKZ8cb4usAVGo17mO0u/CUIVq+fDl+fn706dPnrE1pby2KonD33X/Fav0cmAiA230hDQ3+PPbYv/nxx6/OrIFeOh+CgCCAu9kz142nuvr+/Hw27yjE2nh+HZqtK8qmtqAY2ZUFeIppy67J1BZ1o7Zw1zkVkvQ6qA7g6ptu4rbvv+cWi4WDgh5rgU0OB1tuvpk0jYZKScInMpLvFywgJSXlTJp7WqiurqasrBiY0KRdUa5i+XKvzLuXo1EJAkODwnm+upz/4PFSCvCcWkWAj4lRQ0aQqtGSabfjkzyUmKl3IYhnb2p7S6ktzEJyX8FB5+RBi+SaTm1h1jnloLwhvg5g1KhRXHXnnXQ3GLhbp2OGycRkvR4RWGmzsdZsZq/Fwm25uVw2fjyyLJ9pkzscHx8fBEGhuZIu7Cc4uHUSKl7OH25M68scvYHBKjUPCAL91GoW63T0rKmlwG5nVYOZApeT6L1rKNg4/0ybe1rQmvwR1blHtas0OWiN55a8j9dBdQCCIPD8q6+ybMsW0l58kbGvv860adN4xOXi4NxGAO5VFMTqatauXXvUGIsXL2bQoHEEBcUybNgkli1bdjpvod3R6/Vcc8116PX3A5YDrWUYjX/lkUfuOpOmeenEBOkMvDJoAmO6DaC+Sw/+MmogdQ4H77jcHFScCgTedDup3rTgqP6S28neFXNY9vYj/PnmA+xa9DlO29mdqBSeNhhBWAHMxbOmVICfQPiTiG5Dz6xx7Yw3xNeBdO3ala5duwLwx7x5xDRbKQlArChSVVXVpP2nn+Zx3XX3YLW+CgxhzZoVTJ48g3nzvmD8+LNXmeydd16hvv4OFi6MQ6frgsORw7333scdd9x2pk3z0olRiSKDQj3SNWERLiRFobl8YSxgd1ibtCmKwpbv/ktdSSiyew5gpCjzDarynmHY7c+iaoVgY2dCozcxYMYjbJ37f7gdjwCg1rnoc/nD51wRW6+DOk2MueQSvv7zT26wWg/JEpQAa51OPh82rMlrH374X1itH3MwmQBuwGYz8MgjT5OZefY6KKPRyNy5X1JSUkJhYSFpaWkEBLRONM/L+Y1Jo6FrRCg/FZdzxRHtXyAQ3Gzvpb40h/rSMmT3ag4+6hT5A5zWsZRlrSW619lb8DUgKoVR975KQ0UBoOAbFo8gnHsBsXPvjjop199wA/VJSVxiMDAX+AAYYTTy2BNPEBJyWPhDkiTy8rYBzR3RJLKytpxGizuOqKgoBg0a5HVOXk6J52fezJ16HU+rVMwH7lOpeV6rJ2bMDU1eZy7LQ1HG0XQeLiC5plBfkn86Te4QBEHELzwBv/DEc9I5gXcFdRSKorB/3z5KsrORJYmI1FS6JCe3WezPYDDw+7p1fDR7Nh9//z1+gYG8f//9TUJ2brebkpISAgLCqatbBlzIYRm4rUREJLTJhpbgcrmorKxEFEVCQ0O9IodeThmz00FmVSk1jbXotXq6hUYTZ/I7eceTMKJ3Gr9/9Q7v/vgHL+zaiy0ilajAAZiCDitYOxprD+w1bQKq4Aj1N1G9CWNQx9fTczXWsDNjM4EhIfSL6N7h1zsX8TqoZmSsW4d90yZSAwMRBYF9S5eyLj+fIePGIR6QyFYUhdraWhRFITAw8FD7yTAajdw/cyb3z5x51N/MZjPr588noLaWpy8czie/3sRu1zPYuAHIxmi8iyefbHtRW0VRqKyspKamFr1eR3R0NBqNBoDCwiKWLNmB0xmEosiYTJlMmtSXsDBvlp2X1mF2OliwN5Mxbhfd1BoqHXZ+ra/BEp9Gt4DDzsLmdlHlsGFSawjSGVo8frcucbz+5htsrhX4bEMBa5btOPS3mj0bMGxdyuWKTKVSzGqGkMO7wGjgU0TVEqJ6vtzme3Q77ZhLc3E5rJiCIjEFRyMIArLkJn/jMkpyLXxaokVRfmfL4CAeu8TrpFqL10EdQX19PTVbtzImLu6Q0wny9WVNTg7lvXoRGRlJbW0tW5YuRV1VhQjY/PxIHzeO8PDwNl07c+VK0mw24mJjGXDzDYQbjbwz5z42yA+gNxr45z8f5/bb/9Kma0iSxPLlG8nOdiII4UANfn57mDLlAjQaDQsX7iQgYDgGg6cGRkNDDQsWbOC668YecmJevLSEzKoSxrpdjNN7nE4MaiIlN/8tziPVLwhRENhYUUxueQFxClSgoPMLZExcKnrVqT+W7OZqTFuXcq/JH6NKjTvNj377s/ivdRJ5gg7f0GR6Xfw4WoPvyQc7AdbaMvYsX4nbkQ50AbYSGJtF4uAxVOzJoGp/JGrTVYRGJKIoCps3f8dXqtVcfG7lMHQ4Xgd1BLW1tYRBkxWRIAiEq1TUVFYSFhbGxgUL6C1JRMR41EJrGxvZMH8+I6+5BoOh5TPAI7Hb7VgLCoiNjj50zauvns6FE8ayTlGYfN11qNtB9nrfvv3s2iUSEzMMl8uFRqOhrq6M5csz2bVrHa++uoX6+o+Ii0villtmkJ6eTlFRCGVlZa1WGfZyflPTUEc3ddNJTYRKjdFuo8HtpNxmxVy6nyd1enxEEVlR+NVcy8qiPMbHp57ydRtKc5iIgvGAk1NrdSSn9mFcXQUbB0wmLLl/W24L8EQh9q1bi6LciM4nBRQZQRxDbcFHmEK2sGf5H9jqrwNhPRmOcnr07UtExKUsXLWAyeM7XtHgXMLroI5Ar9dTckQtK7fbTeH+/azfuhVHXR3F+fmYqquJSDysFBro40NUXR3FRUUkt6UixDFKHqlUKgJ8fdvknOrq6pg9+2NWr85ArTaRlnYDubmZOBwq6usr8PWVKSlZxKJFP+J0/hcYQ15eBs888wpPPfUoAQEaJOncK0LppWPRaXVUOe1EH3jEWFwuiq0NZDvtWHN2YJdlblep8TkwGRQFgUk6PWvqqrBFd8HQhs/8scrRCaIKjc54ymMCmMv2Ubh1Bbb6WhwNwQgqH1zWAhREBMGOxtCVnQuew22PAqaBEkBR/lZqqhYz/YopuNwK8jlWK6+j8TqoIwgLC2NnWBj7ysuJDw1l99atWPbtw9/fn5EJCazfto2iwkLk+Pgmqyy9KOJ0OE75unq9HlN8PAWlpcSHek54KIpCbm0tUQMHnvK4xcXF9O8/ArN5MDbbRFSqX4EvuPzy+1mzZgvl5ZWAG0laAMzCIxWvB0bgdDr57LP/8eCDIwgNTT5lG7ycn3QNjeYXcy2RkhsfBQrqq9mqyFxi8GGoIvHv+mokgwmO0CnTAAYUnLKE4RQfTb6RyazJ/J1+khvTgVVUrcvBVlGFb9ip63MUZf5J1pLvkaX7QDEA80BQoTWl4mw0A07s5p14kppuArYCU1Hk4Ths89ix4ycuGhCOWnVuZtt1FF4HdQSiKDL4oovYunIlm7Zto3L3bronJTGsZ0/8jEaGpKTw5fbtVFRWEnFgz0mWZYoliR5t3INKHz6cdfPnU15UhA9QoSjou3enS/KpO4cnnniG6urpuN3/AUCSDEAtP/64CEXpiSxfC+wCZgNjgR3ASiAeMFFQsIiRI2/EZPIGzr20jgQffyxxqbxWsg97Qx1aRWaQwYeLTb7oBIGxegNrrI30NPoeCh7sk9y4tHp8Nad+gFbvF0xNv4m8uWURAxQFtwAbRRUMvgz1Ka6g3A4bWUu+RHZvANKAQmApKFU4G0OAOECL5/BIJHAZ8ApQDqThdm/Cbt/IbVfcApvPjaMipwuvg2qGyWRi2KRJ7IyNxarVMjAp6dDffAwGYpKSWJafz1BRRCWK7LdYMPXu3eQs06ng6+vLmCuvpLS0FJvNRs/AwDaP+euvC3G7j5T38AfCkaR3gaF4UnBr8aycsoDeeGrllQJrSE3V0a3bqe8HeDm/6REYSlf/YL7ZtYn7BIW4I/akJht9udthw9dmoY9GTbkks1QQGBifhtjGCv9Byf1xRCazsmK/56xQRCIavc/JOx6HuuLdiKpeyO60Ay31wBBgIZ7vUG8gE8+x0gYgCHgST4nofWh1C7h/5l1EhQVT2ob7Oh/xOqjjEBISwt5jyKSHRkcTNXgwlVYrsiQRn5REdHR0u8hmqFQqYg4kX5wKTqeTiooKQkND0el0mEy+VFfXHPGKYMCIpxaeHx6H1QXP6ulG4DOgB7Adg+E1Xnjh/BaF89J2VKJImI8flQ01xHHYQUkoRPgGYomIY4HFjF6rZ0xQOCH6tu0THURn8keXmH7K/V12C7LbidYUgEprQFFq8NS8E/AUVqoHrgdWAwHAbcDvwI/A34HHgIGI4nL8A8zMmDED8ta37abOQ7wO6jiEhYWRHRvLzqIiUsLDEQSBvIoKLBERjEpP71QHWBVF4aWXXuP5519CktQIgpMHHpjJHXdcz3PP/QOb7WfAACQjijMRBAFJMuNxUluBxxHFa9FoRuJyNRATk8orr7zBRRdddEbvy8u5Qc+wGH4y12B0OUlTayiXJb5zOukRnUj/0GgIjT7TJh7C0VjLtl8/prZwOwIa9H6h9Jx8A2ptA5LzCzwTOX88E7ln8RymTwE2APvw7E2NxRPig4EjLuLRF35Ep9Md83peTozXQR0HQRAYPGECuzIyWLpzJ4qiENGzJ0P69etUzglg1qzZPPPMZ1itK/HEyPfx+uvX8PjjlzFlShTz5yegVg9HlreRlBTKpZdO4cUXr8Xluhjoh17/Nf7+hWRk7CQ4OBhBEKivr8dsNuPn1/aT/2cCWZax2+3odLpO936db0QZfRjYpQffl+ZTa23AoNHSNSaZ3sFt27dtbxRFYeM3r2CtvRRFXoqCAWvtD2z+7k76TLuP7b88huR6HYhAdq8msucQavK/xtHwDYo8GkhBpbmOyB7D6TbheopqrTx+2wVYGhpobGzkbBXCcDvtKLIbtc502gVWz2oHpSgK69atIycnh/T0dHr37t2qvvn791O4Ywduu52wlBSSu3VrMtPR6XT0HTyYvoMHd4T57cZzz72O1ToLj3MCSMRqnc1rr02kurqIvXv3kpmZSULCYwwYMABBEJg27WJef/199u2by9ixg7j33lcIDg5m795cVqzYi9vthyxbSUjQc+GFA9Dr9Scy4SgURWH16tWUlpYyePDg03qOas+eXNauzcFmU6HTuRk4MIEePdLOG/XiU8HmdpNRU46CQnpgGD6tSFRocDnZWllCpbkajVpDSkgUaf7BTf7f8T7+xKf0RlGUTvs+1BVlYzeDIr/E4RJjVyJLf1JbsIdR975ObcFOXHYLgbHT0Zn8kWWJ0p2rKN25GUG1ldj0SwhNGYDkclC5ZRkvProNUQzja+M33DAyhMGnkGbusluoyd+BqNYSHN8TUX16Ds27HVYKtqyhtrAG0GDwVxE/cCCmoMjTcn04ix1UTU0Nl4wdS+XevfQXBB6XZQaNGMHX8+a1aDm9IyMD85o1dAsMRKtWU7B2Lavz8hgxdepZVzWhrCwfz0btkXSnrq4Mt9tNamoqqalNkx3S09P59NP3mrRVVFSwZEk+YWGj0OkMKIpCYeFeVq7cwvjxLdeZKSgoYOzYSygrkxDFFJzOu/jLX27hrbde7vCHU35+AYsXFxEePpygIBNOp50//9yMWp1D167nvnLxqbCxqpS3d25ggCCgBt5XZG5P68fIiJMfKrW6XSzI2cZIp4MrNRrq7C7m78/GHJnABeFH76d2VucEYDNXgdCTw87JgyL3xVLzHaKoIjih6fdMFFVE9xp1VGX0gi1rsFWkE5pyF4Ig4q+2MfvHf2HoU9sqm4q2LSNr8eeI4iAUGhGEWfSbPpPA2G6ndI8tRVEUclf/QUPVIPS+YwA1Dms2e5Z9To+LJrW5EkdLOWuT8h+44w767NxJtsXC142N5FmtuFas4IV///ukfa1WKyUbNzIoJoYQPz/8jEZ6xsTgX1pKYUHBabC+fenVayDQXKxtMYmJvVrlbHfvLkCvT0Z3oCaaIAhERKSQm2vBYrGcpPdhrrjiRvbtu4rGxm2YzXOx23P49NM/+Prrr1s8xqmyeXMeQUG90B/QxdFq9YSHp7N5874Ov/bZiNnp4K2dG1gkSyyV3PwmuVkjy3y8ewsVdutJ+++oqeACp4NJegPhKjVpGi136PTklhdil9yn4Q7aD/+ILijyCsDWpF3U/ExQXEKLx3E7bNQW1qA2TThUZVyr9cPXOIXFWTUn6X2YxqpCshZ/g+zeiNu5BMm5FrfjSzZ//waS69TPXbYEW105jVVa9L4TEAQNgiCgNXRDdg+jJj+7Q699JGelg3I6nfzw88/82+U6NNfRAs/YbHwxe/ZJ+9fX1xMEqJvtTYQbjdSVlbW7vR3Nq68+hcEwE895phzgC4zGW3njjWdaNY7V6kKrbVquSRAEBEGH0+ls0Rj5+fns2pWFJP2NwzPRQCyWJ3nrrc9aZc+pUFdnw9BsdqfXm2hocCE3E4z0AmurSpgADDqirSdwtQKrygtP2r+2sZ5uzb5HPqJIDArVDttxenVOTMHRhCb1RlRPBFYA2xDEe9DqtxHVc2SLx5HcTlAMCELTMKlW44/Z1vLPYPH2VcjSXzgcugePRlw6lbkde57KZW8EIeqoFa+gisTRePre17PSQbndbmRFofnJhgDAYreftL/D4eDjr77i1muv5bbrr+ez2bOx2Ww0OBwYzsKkgNGjR7N06U9ceOEvhIVNZNiwr/nlly+ZOnVqq8ZJSAjBbC5q0ma1NmAw2PD3b9kWb0NDAyqVH0dHj4Mxmzteajs2NoC6uvImbfX1lURG+rS46vz5hEOSCDzGvkigIuNoQYmrOtnNqppKlpfms7GimHKbFbeiUKEo+JyFirW9L72D5BFdMATchtZ0MTHpRQy55Z+otS2vs6k1+qI1OZCcTTWnas1b6B/f8r1ct8MBSnPtYEAJQXKe/DnXFgz+oaDsQZGbrtQU9w58QjtequQgZ+UelNFoZHB6Op9t3sztR7S/r1IxZcqUE/Z1uVxMv+giEvbu5S5JIhH4ccECHs3M5KInn2Rkly4dantbsNvt1NXVodfrjxL7Gzp0KH/8Ma9N43fpkkB29moKCjbj6xuN02nF7c5l8uTuLX64d+vWDb3eSWPjcuBgXF5Bp5vN9Okdn7bev38a+fkbKS934+cXQmNjLU5nFuPHtzyB5nyif3A4/8zbwfPKYcUkM/ClqGJmcMQJ++6sq2JBUR4TFZnLgUC3i2W1FeQY/fCJiMFf2zlTqxVFwVZXjttpwxgQ3qTChCiqSBx0MYmDLj7l8QVBJK5/X3b8MZvKchtabQS26h3EhK9nXNc4ciqLTj4IEJbSm9KdnyC5ZgIH/5dlKPJighNeOmX7WoLW6E9Yahhl2bPR6CciiEZctg0YA3MJiJ7codc+kk7toBwOB8/+6198MmsW9VYr40eO5Pk336Rr16688dFHTBgxgk0uFwPsdpYajWz09WXFf/5zwjF//vlnDMXFfCVJbAP2AD3cbr4sL2dUYGCHlvVxOBz85/nn+eLDD7Ha7Vx8ySX864UXiIw8eVbMrsxMCtatI0hRaJRlNAkJDBwzptXZdSdCo9EwefIw8vMLKCgowGTSkpo6gMDAwBaPoVKp+Oyz97jyyuk4nbfjdqdgNM4lKqqQhx56q91sPR5BQUFMnz6I7dtzKC3NIzHRh969+xMcfPpmfZ2RAouZb/ZksrW+Gj+VmjHRiUxL6Ea00Zdx0Un0K87jbllCA7wvqugTHkOqf9AJx5ybu503FZk+wPeAHU9xn9/tFl6L7tj6jTtz9/Pso/9hxZr1aI3+yGljiRk65aRJGE5bA3Wr5xBRU0qwIJCLgKvXSILShrRrAod/RCKRI9SM6FFMdUU2k4anMjZyGg3LV7V4jJAufQiKW0VNwQVIrruABkTNWyQOuhi9X8d/nmPSh2IMzKIy5xskp0RoUjhhyRNRncaVcad2UH+55hrqfvuNJTYb4cCnS5dy4eDBbN61i/T0dLbt3cvHH37I+p07GT5kCLNuuumkoaiMzZsZ39CADhgIuAAJGOt2k5eX16H3c+1ll+FcvpxvbDb8gfe/+oqRS5awJTsbX9/jZ8UUFRVRtXIlY6Oj0Ryo8ry7sJCtq1czeOzYdrVRo9GQnJxEcnLSyV+MZzaanZ2NSqUiJSUFQRCYPHkyGRmreffd2eTn/8GECVO48cYbTltNv4CAAEaMGHBarnU2UGm38tTmZfxDcjMfKHE7eaAwhw9sjdzbYxDXJPciPSSSNWUFyIrMjeFx9A48RmipGXmWBiYAEUAvoA7PPD9MkZHpuKrduSXlTLz/3/zNZud1RSG3sYF7ar4h31lHwoXXn7Bv/fqfubS2jAG+QQiCgFVy82nmH9QEROAX0b7RE41PMJdc44no9I/0QZW1ghMFuZ02C+byfej9QtAafBEEkb7T7qVi7yZKs35ApdESk34ngTFd29XO4yEIIsHxPQiO73FarncsOq2Dys3NZfHCheTb7RyM/j6kKOTZ7Xzw9ts8/fzzhIeH8/iTT7Zq3C7JyfxsMsGBrDTNgZ8tBgP3dWB4LzMzkw0rVpBrs3Fw/vGK201eXR1ffP4599x773H7Fu7cSaqf3yHnBJASEcHivXuxDxvWrquogyiKQn5+PhqNhujoY5/0X79+PVdddQvV1VYURSIqKoQffviM3r17k5qayhtvnHg16+X0sKgwhxtkiQcO/B4I/CRLxFSVUmm3Eqo30j0ghO4Brav9GKkzsMXawGQ86TCBeCISOlGFrg2igyfj7W/nc4fDwcMH9s5igaVuJymbFxA95HI0+mNPhByWenwr8ul/wDkBGFVqxqq1fJWXAe3soA5it1nZu7eUGOuxkwsUReH9/7zENx/NAiJRpBIiug2lx0U3Iqo0hKcNIjxt0DH7nut02l3jrKwsBmi1NN+aHOVwsGPjxlMe96qrrmKLwcDrgoAdsALPqlQU+/m1OqmgNXz66aeo7HYSgAnAsgPt461WMtetO2Ffl92Otpk+jiiKaPAkjLQ369evJykpne7dh5CcnE6/fiPJzc1t8prq6mrGj59KQcGzWCz7sFrzycl5gNGjL8JqPXl6spfTR3FDLRc2S4QwAemCSJH11BNXpiR2415RxaYDv+8FrhdVXBybgqqDzjvZ3G4Wrc3gc0kmAfg/oBqIAqJFNda642fhym4nRkE4qhitQaVG6ICMQ0VRmPXKy1zcvw9Dh1xG1JgbePrX1UjN3otZs2Yz59OlSM6dSM4sZGk/Zdkyu//4rt1tOtvotA4qNTWVLS4XzZOb12q1pPXpc8rj+vj48PuaNfw2ZAhBajWhGg0bRo5k6Zo1aI9RHLYt2Gw2dmZm8uKjjzLn7bd5WpbZgKea1ww8pSUz9HoSu554yR6WnExhXV2TtiqzGYKC2j1sVlFRwbhxU9m37x/YbCXY7aVkZl7BiBETcblch173zTff4HZPAK7AM38WgZtwudL58ccf29UmL20jwsef1c0eynZghyITZTj1Kt/DwmKYmtybqRotfoLIIJWarnEpXJbQ/iGo/MZ6luRnc+cPi4lssPANnlriTmAMnhr8JZIbve/xV4E63yAqdEbKj3BGiqKw1WGF2Pa3uWbH73z38VLstgwaG3OwO7L5ZqPEd/tzmrzutVffx277L3AwUhGI7P6QoszfkeXzWyz0tIb47HY7y5YtQ1EURo8efUKJ9NTUVIaNHMkNy5fzit1OGPA58KVWy6b772+THSkpKSxavZrGxkYEQeiQvRGr1crqefOIqqtjz+zZfOh2o8LzGL8ez7mtB4AKjYZtt9123HFkWcbH358dej3WvDxi/fxodDrJV6noO358u5/M//zzL3C7pwJXHmjRIMsP0Ng4l4ULF3LJJZcAUFJShs2WwuEKzx6czhTKzsKzZGcb5TYL+xrrCNOb6OIbcMLXToxN4fGyfLpJEtfhEVP5P1GkV2AY4Ya2ffbHRidyYVQCVrcLg0qNqgNS+bdVl5NflMNwt4tEm50QPM7pMeBdYBwwTlARljoQnen4e9CS044joTcfbF/GOIeFIJWG7W4nW0NiCUlo/yzP6q1/4rbNwaMXBRCJzfUJvxQOYkbC4SSS6upyIJmm36UYFNmF7HIi6lqe4n6ucdoc1MKFC7np6qvpduCBeqMs8+n//nfCtPAv5s7liYcfptdnn9HocHDhwIEseu894uJOXoKlJfj4nPrs8Ujcbje5ubkEBgYSFhYGQE5WFrH19SSFh1NTW8tEPJvIm/FImk0AbhIE1q5YQfhxxA7NZjO//baB6moDitKD3Q27SAvSkdy7N31iYtqsF3Us9u8vxm4/WEbFASwGzDid8RQVHU6P7dOnN3r9c9jtffCo8CYAcWg0vzJs2FftbpcXD5Ki8GHWJtZXFjNYENmhKASZfHkkffhxhf4iDCae6DOSt/ZkcHtDHUZRxbjIeO5J6tUuNomC0KrafSei0eWkzukg3NQuWuMAACAASURBVGBEI6pwShI7SvJ4VKOlwe0iCRiBRxjmT+BiPEdXnw+IYOCU4+/jVuzNoGhrDgrdkITh7HNsJSIuBjEymYCo5A7JTJPs1cDB71IhsBwIwCo1HJJ+l2WZLl1S2L79TjyBygQ8god56H2jUWnbf3/5bOK0OKjKykqunz6dX6xWDlZ0WwtcfNVV7MrLO+4D2mg08vp77/Hau+8iy3KnrEo95/vveejuu1Hb7dS63YweOZKP/vc/avbvp19gIGq1mgCjkXyLhYQDfWzAFiAtMZE+xwlXKorC0qWbsFi6HtKICgtLJ3PjJ8hVK6g2GJBMJlKHDSOhHZM7RowYxGefvUNj4yg8X5QUIASHYwE7dnjSjhsaGjCbfYmJuZD8/C24XBOBleh0HzF6dH8GDWrbhq7D4WDLlh2sXp2Fw+GgT584xo4d0SHJIGcbCwr30lhZQoEs44OMBNzfWM9HWZt5oPeQ4/ZL9gvkXwPGICkKIp2vJp5Tkvg4ewurK4sJEUXqgCsTu9M3OIJIRSFIpUJSqykSBFAU+gNLD/Rdq9Vj6D7muE7GUlNCYUYpWtPfEVWebFlb/Z/U732ZuPI83FuXYA6Nw2/gFHQ+LT9ScTJ0wSnYK+cBu4H38Kz19iNiosjawADgu0VrMRrHIqpCkKWheHSm/o6g+pNuE+5p8/vUWFlI/uaNmMvK0Og1xKT3JrzrBZ3u/T8ep2UP6vvvv2eyonBkudEhwMWKwnffnXwjUBCETumcNm7cyH0338y31dXkWSwUOxzELl/OtZdeis7HB6vDgSAIXDZ9Ov/VatkHuIGdwD1GIw//61/HHbuuro6KCpGQkMMFN8uLsggpqiTVKjE2OpohWi15CxdSWtp+Op2XX3458fEOPGu8t/Gkc8wBcpk161tef/11srPzkOVkXnjhJa6/vh/x8fNITNzDFVeM4rvvPm3Th19RFH7+eRkff7yfPXvSyc+/kC++sPHssx/icHRs/bGzgeXFeTwvS4eqqKiAFxSFDTXl2FqQMKMShE75cPpk9xb0VcXkKzJ5kpu1kpuleTvZWV9FraIgKwohOiMOlYqNQBkehbO3BYFlGi3+KccvZly9PxeECw85J0W2E1D7MzdYnTykMfCobxBXVRdTv+xr5HasHxgxdCpqzX3AN3ic1LfAemTe4rEtm8ivrmXOkr2kpDzI6EkzCIioQ62rQWPsR2yfCYQm9W3T9S01JexYsIyKvaOwN75CQ+WD7FqcT87Ks2eP+LSsoMxmM6FHbLAfJMzpxGw2nw4TOoT3Xn2Vv9psHJy3GoFXnU7iN2/GbTSS1dCAn9HIpVdcgcPp5M65c9ksSRh8fXnymWe44cYbjzu2JEkc+fZIkkRD3hZ6B0YiHjhN4Ws00t3Xl7zMzBYd9m0JWq2WF1/8B9OmPYHTefkRf4lEkh7hyb/+k9ikH8jJDUaWBXr1GsBjj80kOjqakpKVOJ1OjMZTV0UtLy9nzZpKtNoxBAVFARAe3oW9e79n+fLVTJgwpo13eHZjldw0P6Hki+eT4pQlDJ335MhxsbldrKwoZp8ic3D9kga8LEs8U5THBSER/GauZYJOz/C4SJbX1vNKrZlslYpePXoz9sYnyNp7/GLGkltGEA+vvp3WbQyX6klW+wIKoiDQyyeA7eZqcsr3ERDVPlXvjRGpdOs9gO2br+VwnQ6Am3FL/2DKG58ga4dhsc1BrdYTGJNIeNf+SK4oFHlvm69fvC0DW/1Y1LorEEQdkIwsxVOUeRsx6VUY/Nt/i6C9OS0rqAkTJjBHq21ySK0R+E6nY+LEiafDhA6hJD+frs1SRjVAklaLJEnEjR/PioYGVpaW4j9qFM8sWsSuoiLyKyu56wTnngACAwMxGi1YrR4HLkku1G4XsquOyMjDm+I+ej32DnDyen3YMVqDUEsiu/cEIkl/R1G+Ydu2HjzyyBPU1FShUlna5JzAU8i3vFxLQEDTsK+vbxqZmfvbNPa5QHpQBB81k4P4EYjQG/Frp32g002j24VJEGj+uOwK1DjtXBibwna/IJ522Pmv08XGXmm88O7z7M/dw2s/LiYwOvGE4wdGRyK716EonkKtsrueKNmOIDQ2KXMUhYLb0b5HJESVBo6SKhTQ4oNFUmGxhQLTcLsvoaqgntrCPbgd+zAGtV3OwlxWjiD0POCcDtoTjSJHUle8u83jnw5Oy3SrX79+XHLNNQz+3/+4x2JBAN4zmZh85ZUMGHD2nvgfMm4c8zIymHJE6KkE2OFwkJ6eTmBgIPFdutDY2Iher2/VHopKpWLs2F7Mn7+Wurp4RFFHhb2cQZFCk7I95fX1BPTs2Z63xYgRI3C5bsATlggEvgSK0TIHT4nKr/DsIlYCU3E6s/nxx3d56qlpbU7V91TUOFqSQJar2+z8zgWmJ/XgyZpySiQXk2WZDEHgY0Hkka79OmXoriUEafUgimySJY58GvwEpPoHY1BrmJTYDbPLQWiExOCbrsAUk4DbzxdaIK/kH5lEYGw+tYVvIYgXILkK2SrnMzwk+lCNSVlR2K6Awf9YE7NTZ9zUcWRnvo/DcTkeWfifgRrs7Efh70Bf4EPgGhR5BNaalzEGZBOeNq7N19b56GioKIQj/quKXIEgOBHPksnMaYsHvPXhhyy4/HLmfPYZiqLwwk03nbSwa2fnnpkzuWDWLB6oreVal4tC4J8mEw8/9NCh+nVqtfqowq4tJTo6imuv9WP//kJsNjN9+0ymZsMGSmpq8DMaKa+vJ0+nY2g7Oyh/f3/efvt1br+tP7KiAi4BeuDCF88umhqPQEM2sB2Xqw5J2k96+j9OOnZjYyM5OfnU19uIjAwgMTG+iWZVREQEvXqp2L59OfHxwxAEgbq6HFSqPIYOvaBd7/NsJFhn4D+DxrO0ZB8f1VcRZPDlhZgkItqYLn4mUYkiM5J7c+merTwvS/QCfkbgLZWKZ7p0P/Q6P42OKB+aVFRpCYKoosuQsTR02U992UZUGjVVpf1ZWFXIBQ4rKAqr7RbK43oSGnjiArmtZfKV1/DLt3PZsyMOT8DqZsCAp+phAzAVz7HpD4E6BHExMemXYfA7cfhNURQayvdRU1CAIAoExSXgExrXZJIS138wlXmfIznjEDXpKHIxkusLdL5WAiI7tlZieyEorZAgHjBggLJp06aTv/A8orS0lJefe44/FywgOCSE2x9+mKuuugq32817777L9x99hCRJXHHTTdw3c2abM9EqKyvJ274dW20tATExJPfo0W7p8keiKAohIQnU1LyO5zAugAxMxlOl/PFDrzUYruPZZwfy0EMPHD1QM9vnzduCLCeg1fpitZYTFlbDxRc3LddUV1fH++/PYccOK4JgJCJCZMSIBC66aMQZVzsWBGGzoihtWvYn+wUqrw44v/fSjsXWmnIW5e+mym6li38wlyZ0I8roQ6mtkZ/3ZZNXX01soJEH/+9mxk8cjzu+D5trBT7bUMCaZTuIDWz5CluWXNTuy4R920AQoEsfAhN6I4rtl4xVWGtl1mNT2LJmJQ/ccA+ynMXhcF8ekI5Hv+1gOLsKUdWFUff994SKtYqiULBlBZU5IKpGADKytIzI7iaiezXN5CzevpycVVuQnCZElR6dr5WkYUMJSz6z1f0XvXhli75HZ9+OaicjMjKS195+u0mboihcPXUq5pUredJqRQ288dRTLPrpJxatWtUmXaLQ0FBCx3T8w23v3r3Y7QpwZKKECPwNQbgZRbka8EcU38dkWsktt7x97IGOYOXKHeh0fQkI8IRRgoOjKCraSVZWDn37Hl4FBgQE8Le/3UpZWRkWiwV/f39CQ0PP2hCWl5bRJyicPkFN9x6LrQ38c9Of3C25eQrYUWrhjn/9lxckkel3n3pFGVGlITh5ACR3/BbDisVLUJS7aLoX1YXDE703gVzU2vuI6HHhSeXUrbVlVOba0Pv9FUHwTNgUOZ2yrOcJTqhG73t4CyC61yjCUvpjLtsHKPiGxaM1tkzbrTPgdVAdwNq1a9m+ahU7rdZDhWFH22wM2L6dRYsWcdFF7a+LVFZWxt5Nm2isqMAnLIyUAQOIiDj1cIVGo0FRXHhWTUfOKh0EBWlxOgfjcFgYP/5i3nxz2UklOex2O+XlLmJimsb4g4JiycnZTN9mGbWiKBIVFXXK9ns5N/gxbxcPSm6eOPD7BUAPh5OrXn6Py+94gPZ+hMluFzV71kPOFpBcKHE9COgxHI3+1KMUGq0GQXDQPFil17iJDl9GflkIJr9wuk+8HN/Y459lO0hjVREoAw45J8CTCKH0w1Jd3MRBAWj0PgQntM+h7NNNp63F15mRZZn58+dzz6238sjMmWRkZDT5+5o1a7jY6eTIbUgVcGljI6tXrmx3e0pLS9kxdy6pNTWMCwwktaaGHXPntul8VGJiIomJcQjCh0e02jEaX+Lppx/CbK7A4bDw66/f0qUFB4VVKhUqlYTU7JyJy+XAYDizYTsvZw6zy8G8/N28v2sjvxTupdHVtPrmnvqqQwHmgwwCXHYHpaXtW1JLURSq1s9j4LY/eVCAv2l0TMrdQs2fX3pk3E+RCZddhk73CVB8ROsmYDVrv36VhsZ65m/ewqDptyK0IMToOZB8jCK/Qj3iWahifCK8DqqVyLLM9Vdcwd9nzCDlk08wvfMOk4cN4923DovxRUZGskd3tJroHqORqJiYo9rbyp4NG0j38yM8MBCNWk14YCDpfn7s2bChTeP+8MNnhIa+iK/vGPT6uzEa0xg/Poo777yj1WNpNBq6dw+lrGw3B/c93W4XdXW76dUrtk12ejk7KbY28PC6xTj3Z3FpeSHmvF08sn4x5bbDZ5oCtXr2NOtXBVhkmYCA9g1V2eorCC3ezRT/UAI1OnzUGob5BdPPXE198amfS0rp3osnnngAvbYnes216MWpaMXRvHX1eAJ8W78y849KRqXehNtReKjNZc9FpdmJX/iJU+7PNrwhvlYyf/58di1dynqLBT2ALHOzzUbfv/2NK2fMIDQ0lMsvv5xHZ87kI+AWPOUfvwOWiiJPjx9PRUUFwcHBraqOcfChfqx9mIbycoIPlIuqraujsrIWAYXSioo23WvXrl0pKNjNr7/+SklJCUOH3kb//v1PebwLLuiNzbaJnJzfEUVfoJbhw2Pbrbail7OLL3Zn8JjbxUMHfr9Llnhalvhm77ZDZZvGx6fxSNZmesgSyXgKAd2t1XDp6CHk7ttHudh6ZVlFkRGEo+fmjoYaeh6Q45AlCYelDrfDSaxkZ0tVIbRBuO/hR/6PGf2i+faDzyjbXc0FIRcysvupZdJp9CaSR1xA3tp3sJujARmNoYykYSNQaY6eGJ/NeB1UK5k/Zw43H3ROB0gELtRo+P3335kxYwZGo5Hfli/npmnT+EdhISpBwCc4mPtue5ylSysQhEpMpgzGjetNeHj4CZMmampquP/+vzFnzv+QJCdjxlzM+++/0iSsZgoOps5ioaq0kr17LajVYTTYG9gimOm1czc9eqSd8v3qdDqmTZt2yv2PRKvVMn78UAYNMmOz2fD3T/fW1ztPkRWFTXVV/N6s/W7g5ZryQ78PD4uh1mZl4P4sQgSBClmie3wSvg2x/Pu5rdQ4SnAEgeSfjKIYTphIU71/O1lL52CpykKtCyJ+wASShl16KKymNQWQLyu4XU7MJUVI7mAEIZjdriIKsvfinzrwqP2d1hAXGcb1g/qQYc49+YtPgm9YPL0ujsZaWwaCgDFwaLtmIHYWvA6qlRhMJsyiCLLcpN0sCE3kQ3r27Mmm7Gzy8vKw2+2sX1+ATncBfn7BVFdX8/rrb3HTTbcjiuVcdNHlfPDBa0clBXhkSaawe3cfnM5cwMjvv7/N4MEXkpOzHT8/PwCSBw5k1TffYNttJj5yII0OG0UuO8n9rmfVqlwSE2M71SFXPz+/Q7Z7OT8RAI0g0qDIHPnJNAP6ZhO2qfGpTIhJoszWiGSy8Gv9IKLD70P0j6YiZz/bf30LyfVXcvUGugyeQsKgyUc5qvrSHLbMeRvZPQu4FLdjD/vW347L/i3dxl8LgDEwgtLwBBbu2UwfVxw6bRRZ7mo26LshaqZStHUtySMmdOj/pTWIKjU+Ie2/ZdCZ8O5BtZLrbr2VD/R6Co5oWwLswFPS6UgEQSApKQkfHx8cjnD8/IJxuVz89a9PkpmZhKJ8iyRtYOHCJAYPHoPT2XQjdtmyZezbZ8HpfBfPWQlfZPlxLJYL+PLLw5IWsXFxGPv1J0Mdyh+1pWxWqfC7YCqRUcnIcgQVbQz1efHS3giCwOiwaB4XBA5O9STg76LIyIijQ746lYp4H382lgkE+E5CrdJTXFxCxrpNSK6HgJtx238ld3UGeWt+Pqp/7uqFyO6ngGl45uXdkd1zKcpceqi8kSAIBA2dxg/qEF4V1LzoquQrYx9sUY+h8x1KfUklynkuIHi68a6gWsmAAQN49JlnSH/iCcaq1dQJAtsFge9/+eW4AowulwtB8MSG169fh8USjCzfAGQBKiTpOerq1jJv3jyuvPLKQ/12796NJA2GZrXXrNahbNuW3aQtNi6OhH6RREZ2axYydKJu5cl7L15OB9en9uE/1gZSLA0MAlYD4b4BPJx0/MooVreCXuWpmrE5Y+cBiYoueB5lXZFc37Fv/RASB09BVB3+3DdWlgLNK56HI4jh2Buq8TlQk0+t1aMOTaNBuR+VJhRBUKECZKkBUS3CMfauvHQc3ifXKTDzoYe4+rrrWLJkCSaTiUmTJp1QHTgsLAxFWY8kpVFUVIzNloZnvlgC9AOgsXEwu3c3LeDYvXt3VKo38ZxFOvzFMJlW0Ldv07NUMTHRqFSrsNvjMBo9B/3M5mqMxhoiItpWtt+Ll47ApNbwVP8L2WOupcjawEyTH8l+Jz5P1zdMYJV5C75BiZgbzEAYkAkE4anr7ociKbjslibqur5h0dj+v737jo+iTh84/pnZvum9d0qAoPReBSlSBM92iu301DsUf6dnBc/eznLqYb9TwXKcDUU9RUCR3kIvAdJ771tnd+b3x0JIgEACgWxk3q+Xr5dOdma+E7PzzHznmeep+xVonuRTjCKXYzyurFB4t3gKdy1HE3A94EmqcDT+SETPOPVl8fNMDVCnYLfbyc/NxVpXR0BYGHHx8U13IxEREcyZM6dN2/H392fYsCg2blxHQIAeg2E1DkccnhL8QYCCj8+vpKXd32K90aNH07NnGHv23IrT+RhgRqNZiK/vLq6//qMWn/Xx8WHq1DRWrFhPTU0QIOPrW89llw1Q76BUnUpRFIqsjeTUVyEASQEhRJt8EY70puoZEEzPgOA2bWtCQgg51WvJL3HiY9bgdH4I5AJ34Jlp2Iuo1aAztUzfThl5GZU5TyO7wvE04jyIqL2T+IGT0epbXlyG9+iPrf5XqvOeRhCSkZV8AmM0RKeNP9tfhaqd1DNXK2pqatiybBnRVivBBgPldjs5kZGMnDYNw0necTqd/v3TiI4u46KL8li2bB1lZU5crieAAnS6vxMVZWH69Okt1hEEgZ9//pb771/AJ58MwuVyMHXqTF5//deT1t+LjY1hzpxwKisrEUWR0NBQr2z0qLqwbCoroLosn1GIgMLa8kIKIhIYFtn+99/8DXr+fudEthTV8E16AR9+9Cku1wtAJLAGje52UkbNPCGjzT8yiUHX3kvGypdpKL8NnTGExKGXkjjkshP2IWq0JA2dQFTvahyN1ejNgzEFHN+FS3U+qBOqrdi7YQNpikLf2Fjiw8IYFBdHRFkZh/fvP+NtRkREMGrUEHbuXM3vf6/Dx6cfPj79ue46iQ0bVp70Tsff35933nmdxsZKtm7dgNFoZMqUa5gz54/s27fvhM/rdDqioqKIiIhQg5Oq01XYrZSVFXCvwcgYo5ExRhP36o0Ul+VT5bCd0TZ9jAYmjxzE6y8/yovvv0Nw/FsIYiRG/5tJnXgJCYOmnHS9oNhUht88n0vvX0yP8VdQdnAvGz94luxN3+KWTuzWbPQLJiCqmxqcOpF6B3USTqcTS0HBCWnf8cHBbDl0iLTjCscpioLVasVkMrWpEGxoaCiLF7/D4sXvtHlM69evZ9KkWdjtf0WWbyEjYy1Ll47j55+/Y+jQoW3ejkp1PuVb6hmIgrFZcoFJFBmkQH5jPSGGltNrblnGpSgY2nhxNWjkWK54Iald1cz3fv8+pQfLkaX5gC+W6n9SduB5ht44v0ViharzqXdQJyGKIooo4j7uXSfJ7UZ7XEO+xR9+SLeoKEIDA4kLCeEfL75Ie1qYHE8+bp9H3X33I1itryHLDwKjkOWHsVpfYt68+Sf9vErlDbSCiI0TEwusAmibXcxZXBJv7dvCnDXLmLNmGY9t/ZmshjZ0I2yFoign/R42VhZSmpGOLK3Bk3I+Gdm1DEu1D+WHzq40mKrjqQHqJLRaLWG9e3Oo9FgxSlmWOVRdTWyz5oCff/45j8+dy8dlZVhdLn6sreWjxx/n1Zdeatf+FEVh4cK3iIxMQaPRkJx8MV9++VWLn+/cuQ648rg1ryQ9veOLz6pUHSXFL5DtgkhxsyLBhS4XO0SRFL9jjTxf2bWe6Ipi8hSZRhTmNdbyzI61VNrb14LdYalj59K3WPHijax4cQ7bv/gn9vqqpp97Wp1PBpo/wxVxS9dQlds12qBfSNQA1Yq+gwdTFR/PLwUFpBcVsbK4GP3AgSSlpDR95sVHH+UNq5XhePKH+gKLrVZeeu65Vu+ETubVVxfy4INvUla2BHCTk/MSN9wwj++//x7wJEv4+4cDx5dIySIwMOL4zalUXsNXp2dwYiqvyTLv2m28a7fxT0VmaEIqZq2nin12Qy1ljXW8p8iEAXrgD8C1sszKopw270uW3Wz+6FnKD/dFkQtQ5DIqskazafFTTdXI9eYABPHEUkOC5hAGv67TJ+lC0WUnXBVFYdOmTVRWVjJs2DDCwjr2QabBYGD0ZZdRXV2NzWajZ0DACZlzmQUFDD5uvT5AdUMDNpsNH5/Tt+GWZZmnnnoeq/VHPCEO4FJstn8yf/7zTJs2DYC77/4Tr7xyF1br53je+ajEbJ7HPff86SyP1Ps4HA7+97//UVZWxujRo+nT58yLdKpOz+KS2F9biUmjpVdgKJoOftcnxT+IuF6DKbDWIyDQ3+yHvtkzphJbI/0FgeOfOg1TZD6w1LV5P5VZ23FaQ1Dkf9D0crvyNC7HZsoyNhGdNobQ5H5otItxO98Gbsdzjb4SQVhCzEXPneWReh9bXQWVObvQ6AyEdxuI1uA9Jc/aoksGqOzsbC6fOBG5ooJ4UeQGh4MHH36Yhx97rMP3FRzc+vsZF6Wm8vP27VzdbNlGICY0tM217xoaGmhsrOdYcDpqBFlZx9paPPbYI5SVVfLRRyno9Uk4nTnceOPNPPLIA20+lq5g3759jBs3FYejGy5XMvAkv/vddBYtevusOhGrTu6nwmw+ztpNf0GkGqjSaLj/4pEk+Qaedt320Gs0pPid/CXcRN8APlBk7NCiCPNPoki8f8virDl5wH+/I+Wa6YgJ/TylhxSQFSgqzMMtjeT4yituaTTFhem4YzzThXHT51Gw/GUkyxMgmBA1NvrPuhvTcS/sdnVZ678he8N3IE5FEGrZ/+Mi+v9uHiGJndvuvT26XIBSFIVrpk3jlrw8/iLLCHjqMYx98UX6DRlyQrdaSZIoLCjA2tCAf3Aw0dHRHZZ+veCFF5gzcybYbFwCbAXmms08/txzbX7j3M/PDz+/QKqrdwDNswPX0KPHsTsHrVbLu+++znPPPUZOTg7JycmnDJ5dkaIoXH759VRWPgbcemSphaVLx3PppR9z4403dubwfnMy62v4ImsP6bJMtyMV8T5xu7h/53oWjpiK5rgLggq7lez6ahQFEv2DiDSdfoagLWLMfvQJCmd2TTnPyzIhwL+AFRotL8YknvD5piAFDBo6HWGop3ZffkANv+xajGRXaB6ktMbV9B01gW4jjz4/TkOZdSl1JXm4JSf7M+1YNSJnXqfc+9QWHyZ74ypk935wRx1ZuoYdX17O+HkLu0xbji4XoPbt20dlQQH/dyQ4AUQDD1ksfPDPf7YIUI2NjWz89ltCamsJ1GopliQyo6MZMXXqGb1se7yJEyfy8bJlPP3gg9yZkUG3+Hj+/tRTXHnl8ckMrRNFkSefnM8DD1yH1fouMAhYicl0D88++8EJnw8JCSEk5My+SitWrOCVV96lpKScyy4by733ziM01HuuGg8cOEBJSQ2eLlpH+WCxPMDbb/9LDVAd7NfiHO4+0mfpqOuBl2U3e2sruTg4vGn5jsoScouyGYFnUmxDaR5RkQkMieiYatp3pQ1jaW4GlxXnYJNdDAiO4MmUvvi3ciJtHqQGDp0OQ+NxD7qWjC8/oLhgLi5pPqBFo32JkOAiHv3zzehP+M4nALBocz7rVu+loMbaplR1l9NGfvoKSjP2oNUbiB84koiew7yqDFLR7g3I7j8DUc2WjgHhYiqzdxLRs2u8mtLlAlRdXR2hGs0J2R3hQF11dYtlezdvppvFQtKRLrbJwL7CQg7t20ffAQM6ZDwTJ05kYnr6WW1j7tw7MRoNPPHEHZSUZNO9+8W8+OJ7TJ48uUPGCPDaa2/wyCMvYbUuABLIyPgvixYNZ9eujV4TpJxOJ6Jo4sTcHTMOx5m33FadnM3l5GQpNuGA1S01/Xet087homwe1BvwP3JXNUKW+XtpHhUBwYQZz/65hk4UuTq5N1cn927XelnNgxQi63/6Lw/Of5aly9KQZTfTpk7jpee+JDJCD5z89Q9lcAxAm4KU2+Vk0+JnsdX2RHY9B1RTX/o0NQU5TW07vIHsdoNysjtcM3KzjEpv1+UC1MCBA8mVZfZw7KmNAnxoMjG52Z2L2+2m6vBhhkRFtVg/MTSUDQcOdFiA6ii33noLt956y+k/eBqSJJGRkUFgYCBxOpaANAAAIABJREFUcZ5SMhaLhUce+RtW6yagOwAOx0Sqqm7j1VcX8vTTj5/1fjtC3759MZnsNDauAC49stSNyfQG118/szOH9pvUNzSa96vK+IPsbkpQKAA2KjLXBxy7aMlrrGcANAUnALMoMgSF3IbaDglQZyInD5ISjgUpfWQsEcCHD83hw4ea1cm0l0BeSavbGZTQD04SpJzWeuwN1ZiDIprq9ZXuX4+tLgrZtZSj04huaQqFO5NJHDrJa55jRab2pyzjbdzSHdDUceswiryOkKSrTrWqV+lyAcpoNPLqm28y6c47uctuJ0GW+Y/ZTHlCAovuvLPpc4IgIGg0yIrS4npclmXE31jxVJfLhSAIfPXVUu644x4kyReXq5r+/fvz5ZeLyM3NRatN5mhwOsrhuIqlS5/0mgCl0Wj4z3/+zeWXX4vLdSUORwq+vl/Qu7eZP/3pjs4e3m/OyPBY1hRlM7axjttkN5XAP0QNVyX1wl9/bDpMIwhIJ1lfAjSd3H6iRZC6ZvrpVziZ0kIGDpyEMigat+Rk3eo9bPz8Qxpz0xE0UShyCUnDZpAy8nIqsg4hS7+nZSJGMIoygpqCA5j6jO6Iwzprocn9COu+hYrDF+GWbkEQaxDED0idOAe9ya+zh9dmXfJMff0NN9CrTx/ef/NN9pSUMH3aNG66+eYWmXOiKBLRuzeH9+yhV4zn6khRFA5VVhIzblwnjfzsKYrCmjVr2Lp1KwaDgeVffslP69ahEQS0soZGeRkwCXCydeuTTJ16JV98sQhJKsRzStE121oOhw8fpKamhqCgU7c5OF8mTJjAwYM7+fDDxRQVFTJx4gPMnDlTrch+DmhFkYf6j2FdeQGLyoswanXcHZ1Mr8CWzziT/QL5RhQZ53YReaQUUKXbzWZBZOpp2mOcD82DVHs0OiVW5RVT53Ri7vkNn/60kbzKSkwaHRZXd2QlD9yBQB65m6diCgjC6OcLQuZxs4UKipxNyf5aor0kQAmCwEUzbqc6fx/lhzaj0RuITvsbviExnT20dumy3/oBAwYw4F//OuVn0gYNYlN1NRUFBQSKIpWyjLlXL7qnpp6nUZ5abW0t+/ZlUVraQECAll69koiJaf0PyGazcemls9i1qwC7fRx61794FIn/Ag5gATKLuAcr+wA9LteTZGamYLFYSElJYO/eB4AX8LwKmQE8hyD044MPFnHvvf93Pg65TWJiYpg//+HOHsYFQSeKjI9MYHxkQqufMWt1DI7vyT/yD9JbktACewSRfvE9WtxpdRZZUVi2o5L0ahd2t0yyr5thYeH4642trpNRV8UTu9JRlFE4ZSfhW7/hS2A4kC47uJIs8vkUhT8DCbilf5Cz6S/0v+IO8tL/hqdlx3A8fd0WAjJVOXtw2hq85g5FEARCEtIISWi9AaS367IBqi0MBgNjpk+nsrISq9VKrJ+f16RmV1VV8dVX23A4oikosFFUVIMsb2XatG7MmDEWf3//E9Z5/vmXSE83YbfvAd5hEnoeOjL54gO8gcwKisjkZ2AiIKLVplBSUsJtt/2ee+99DVn+FE/eYyHwDE6nm40bN/Dzz5upr3eQkBBMamrKKRswdga73Y7dbsfX11e9m+oE3QKCie09mNzGOhQFLvf1b6oE0dl+Ki5lU0USDncE5Y5cNpQrfJmfx+w4I5dGR5yQLu9WFJ7dswub+z/AdHxJ4SOO9dsdBHyOjfE8gYU/H1naHaelGp+QGAy+ATgaZgFhQB2eTLkfEDTTKdqzFnudgqjVoIRHIMtTver9PUVRcDRUI4giep9Ar8o8PJnf/DddEIQOrzLREbZsyUAUU8nLO4TDkUZiYiwNDdVs27YL2MQ111xywol40aLPsNv/BWjQsY9LsLT4uQCMwEUmB/EEqFIcjnQGDRpEUFAQJtNCLJYNQBWeFBMTBsMsZDmN/PxYDAYfNm8uISNjHbNnj8ZobP0K9HxxuVxs3ryLvXsrUBQTer2NUaO606NHyulXVnUoo0ZLaoB3vS1U7bCxpdKEThxEgSULveZRjJpQbNJuVpV+jl5zmEuiIlusc7i+GkkOAzzPrOwUcHzS9SDASgXgwnOa/ILAWM/MS1BsT0oPjAOm4Lk0TAVKkF3VVGQmovcZCbKTxsIf+XbJF1x+3dV4A0t1CTmbNuKwmEFxYw5ykTR0FEZ/7/p/2pz3hPYLTEFBLYqi0NgYSECAp5W02exPeamdXZtLWb5sGVVVVS3WcbvdHL2mkBjId7RMI5WBX5CO/Nt/8fGZwF/+cg+hoaEMGTKEvn0TMBieACIAO4LwNFrtIaZPv5uQkBh8fQOJielFXV00Bw+eWK+sM2zbtodduzRERFxKTMxY/P3HsGJFPsXFxZ09NJUXKLdbEehJuf0QWvFyNEKoJ31BCMPiHMW3ebVsLCvA6jqW5iErCjQrrGSiG6uP2+5awEwAsBWER9HoX6DHuFkAJI+Yiqh9GdgPxAN7EDRTMPhOwhx0PTpDEjpTT7S+f2TtigyqK8rP4W+gbVwOG4d/XYfLcQMG34cx+C3AVn85h9b84tVp5xdUgCosLGTJkiWsXLnyyMm+8wQEGGloqIYjQcbtdpF/YAOBFQWE1Lsx7d/P9s8+Iz8vr2md6667AoPhFTxPaK9lC4Hci0gRkAncpNejjwyjZ8+PGTHiQz788AmeecZT/kkQBH76aSk33hiIj88AtNpoJkzYxkMPPUloaMu3Yfz9o8jPP/NWBx1FkiR27y4lKqpPU/UPg8GMn18v9uzJO83aqnNFURT21Vbya1kBJdbGTh2Lj1YHlOGUbYiCZ6bEKdejcWWR4NYQ6tLhV5LLd4d30yh53qXr4R+MKBTDkbDUwAvchIlleCbslgPX6vS4g0MxB99KdNo2ht/yOL6hnvcp/cLiGfz7+/CPfAWEAHTGSwlNCsIvfHKLKTNB1CMIqRTl556n30braosP43b2R2fq6clwFgQMPgOQrEk0lHf++Frzm5/iA88XasH99/PWwoVcoteTB9T6+fHVjz8SFhaGTqcjODj4vM7HDhyYRE7OfiRJh6L0oqqyAFNlJoP6xmNTMumbmIhJr2f96tXEzJmDRqNhwYIH+eGHSeTkjKCxcSqysR//dlfxgVbAqNfz++uvZ8fzz+Pnd/KHtH5+frz77uu8++7rgOe5zqJFq3G73S3KP9ntjcTFdf70niRJyLIO7XHPOgwGMw0N9k4a1YWt2mHjuR1rEZ02+iCwWJEZHBbD7KTeWFwSIQYTAecxcSLG7EeEqYByezxW9z6M4mhczsP0xkGApgaz2Mh1RjP/c9jYVVnCyKgEtKLI/X368tyeGcjKLCQlAYvox50GaHC7iY+Jo9fMP1Jqj2r1pd3AmJ4Mv/lYL7ayw9sp3FFxwucUpQxf/4vP2fG3leSwoQgnSYRRwnE5as//gNrIqwOUoijkZGdTsHs3ks1GWPfu9EhLa/MDfIfDQUNDA6tWrWLp229z2OEgxOFp7fx0QwN/nDCBV594AifgCg9n8KRJrZ7cO1pKShIzZkgsWrSczMwyHLVOBkRrsMhF9I6TCPX3RxAETNXV1NfXExQUhJ+fH9u3r+XTTz9l9eo19O8/mZtu+vSkCRVtYTQa6dMnlN279xId3QeNRovVWo/TeYhevfp18BG3n8lkIjAQGhtr8PU9ls5cU1PM4MHeO2/ujeqdDnZUFFFWV4VeqyMlNJq0oLA2XZQpikKlw4ZTdvPBgXSutjXyFJ5nnhXAuLJCvm2opb/JzHYFQkOjGBOdhHgeLvhEQeDapHA0ZLO2/BXq3QcwKYHodPVY+IE/mq0IgoF+Wj3b6qshynOS7hccwVtDR/N90RYaXRu4Mq0H1/7fLWiGzSC9RmDRlnxKV+9t8ziC43pQvOcHJFsqWmMPQMZlWUtUnJWElO6nXf9c8w2ORFB2oChjEY68u6YoEgi7MAd5b9kjrw5Qe9LTady0ib7BwRh0OgrS01mfnc2Y2bPRH9fZtjlFUdi/axcFW7bgL8t89tprXGuxcLTbSzUwFMisqyNCkkhKSqKwspKtP/3E+Cuu6LA7KYfDQU1NDQaD4YT3jARBoE+fnjzzTDIZGQdZ8d0agkpLGd07luTIFARB8GTcQNOxOhwO7vrDH/jiq6+I1ev5esmnNFRX8/Df/nbGYx42rB+iuJu9e1ciy3r8/Fxcdlkvryh/JAgCo0f35ttvt2GxdMdk8qOhoQx//yJ69RrZ2cPrMqwuiR8y93CJ5OB6nZ56ycH3+YfY6LAxIqr19HLwBLZVeQfRWRswKQpV9dWM49hrqt8Bt6MQ4bAxLDAEp6LwYUURu4xm+odEtr7hdlAUhXK7FYtLItxoxlfX8rvvrzNwa48kpsQ2sLHsRzaVVNNf5+Yys5sUredurk5xo9cee2ab3VDLG3s3Y3N67sRz60tJnX4JA4ed2Rh1Rl96jBtJ7paPsddrAQljqI7b/rLAK7L4fMPiCIzNoKbwfbT60SiKG7fzZ8K6+2MK8L4ksqO8NkDZbDaKt21jYmws2iPTT6kxMTgKC8nLyaF7z56trpuXm0vNunVMiIlBp9XyoywTD+wDLsZTziUFCNRocBy5o4oNDSWrsJCampoOSUXft+8gGzbkIMvBuN0W4uJEJk4cgslkwul08tlnn7H8yy/xDw7m5jvv5IbbrmL3558THRLsaTmvKBwqKcEnJaWpr9QD8+ZRsXQpuXY7AXY7+cD0v/+d6Ph4br7lzMokabVaRowYwKBBTpxOJ2az2Su+UEfFxERz9dVGDhzIoba2iLS0QHr0GOV1afDebF91OYMlBxONnt9ZCBpuFTU8WV7ExWHRR57jnEhRFFblZTDBZmGUwYjd7Sks+xOeupZxwA7gPmDDkfbqekHgMp2e9yqKOyRAWVwSn+eWkW8JRRQiUZQsxkQojIkIRxAEKuxWVhRlU2GpJ94/hEmxiQRqncRWl5Og8UxT18sy37vcdA/zlD2zuSSe2bGW19wSR6vnLWm0cuVDL7Jr5k3Amc1I+ARH03vyLJzWekSNhlKbSECQd7zWIggiScMnEpC/n+rcrxBEgdDkeAJjenT20E7JawNUfX09QYLQFJyOCjOZKC4thVMFqF27SAsKQnckTXvwmDFszckhTJJIw5M4WoQn2TqlWYdcgygiSScr6tI+JSUlrF5dTFTUeHRHqjGXlmby66/bueSSwUwbPx7Hrl3cZLFQIYrMWrKER//+dyZPmcKaNWvwrazEpiiYk5MZONrzZrrdbufDxYvJsNub7gTjgZesVha88MIZB6ij9Hr9Ke9KO1NwcDAjR3rHF70rqrHWM+K475FJFInH80yptQBVbrdisDYyymBEEASMGi2hopYg2cU64GhFtywgpFnrDbMgInVQEtJ3heUUWCYTYhiJIAi4ZTu/lH5EpKkSUYBnd67jBllmhiKzvKacBwoPM7//WPa53TxZV0UoAsWiQI+YJLod6Ue1vqKIEYrM9c3283vgC1nmy6+/4aIZN5zxeD3JB0e+obb2tas/10RRQ2hiX0ITj+895728NkCZzWYaFAVFUVpMX9U7HJhPU5ZHsloxNjvZTpo8mbUrV/JqcTG5TicHNBp2iyLP3X03Op3ny2mx26k9kixxtjIyCvDx6dEUnAAiIlLIzc1h0aJFOHft4heLxZPoKstcbbUy8K9/5drrr2fiDTdQV1eHwWBo0cG3sbERUVE4/pq0O1BS3vlprCrv5WMwU1RXTe9mccilKJSg0FvX+kWJQ3YTAE3fP0GA1KBQvqkqZSvgBDYiECaKXON3rMHhNslJZGjUSbfZHhaXxKE6A0GG4U1j0IhGTNpL2FL5EQer9vGa28XRcDJHlpkvO1mWs5+5aUOpddqxulwMMxgxaI6d6mocdnrKJwbQ7g4npWVldJ12fr993jOXcxw/Pz98e/RgT1ERksuFoiiUVFeTp9eTkHLqlzRDu3WjqOZYmrTBYGDeo48y4s472X7NNZjvuou577yDEB9PQUUFmSUlbKiqotf48U0B62w4HC602pZffE9qp57lX33FLUeD0xHJwBC9nrVr16LVagkJCTmhvXxISAjhISH8cty+vhIERowYgUrVml7B4fwiiuyXnCiKglWW+cpuwzcwhMBTlAOKMJrJF0Vqmp3Mg/QGHAEh2MJi+Dwshh6JqewJCme508EOp4MlNiu/6o0MCD/7PlGS7AbBiHDcaUojmGmUFDKtDfz+uHX+COyoLgMgUG8k2uzbIjgBpAaE8LWoaVEAVwK+NegZPnTIWY9b1XG89g4KYODo0ew1m1m5Zw+43fhERzN41KimZzKt6ZGWxvqsLByFhYT7+FBnt5Mjity1YEFTVQlZlikuLqY8Lw+dycTg5GQCAzumzXVSUhhZWfkENHv42NhYi6+vk5CwMCoFAZSWvWkqFOWUGYSCIPDCwoVcP2cOj1mt9AOWazS8YTLxywsvdMi4Vb9NwQYTo1LS+LwwC5vNglsQiAuLPmX9PQCDRkuf6GQWFmYyURDwFwW2uNyU+QczN7k3OtFzmVUvOcioqSDLbiPQx4+ZgaEYNWd/avHXGQjUVWF1FeCji29abpF2MTgE1pQK1KG06IRbCZhOs++0wFBCA4KZXFfNA7IbAXhJoyG+RxJjR49iu/dmXV9wvDpA6XQ6+g8bhnvwYNxud5ufkZjNZkbPnk1uVha5JSWYg4MZ0b17iwAgiiKxsbHExnZMR9DmUlKSyMzcQF7eFkymaCTJCuQyY0ZfkpLu5Jovv+Qqq5Wjp4dPgDofH0aPPnUl5FmzZxO+YgWvPfMM72dmMmDYMNYvWED37p2fxqrybjFmP6K7X4xDdqMVRLRtTIS5KCSCEJOZzVVlSC6JyIAQpgWGNAUn8ASSIR1wx3Q8URCYGefLJ9mLqbKPQyOGIrn3EWXewdCwKPZURvFgZQnvKDIawAY8JGoYG510yu0KgsB9F41keVE2D5bkAjClTyIPLbjL62vTXWi8OkAdpdFoWrxI2hZGo5HUPn2gT59zNKrWabVapkwZSUFBAUVF5ZjNelJShuHv709UVBT3P/UU/ebPZ4ReT7miUGU28/Xy5W06xhEjRjDi++/Pw1G0jyRJlJaWEhoaqmbYeamjiQ7tFWP2I8bcORW6E3wD+FNPO7tqVlHrhCRfkV4B0eg1Gm7u2Z9XHDaSGuu4WBDYqMj0C4lgVkLrCVRH6USR6XHdmB7naXiflAB6nXecDiW7Bbdkx+B7fosHeCPv+D9yHtlsNj766GOWLfuZ8PBg5s79AwMHDuzw/Wg0GhITE0lMTDzhZ/PuvZfrb7qJNWvWEBAQwNixY9sdgL3J66+/yd/+9hSSJKIoVm6//XZeeukZter4b1yJtZHviwootkr0CTQxOTrhhHeUOkKQwci4yBNT1n11ev42cBzZDbWU2ixc5htAtNn3JFvoGmwNtaR/9hpVeTsQMKD38SftshsJ6UJZdx3tgjqDWK1Whg+fSGamP1brdYhiEUuWTGfhwuf4wx9uPq9jCQkJYfbs2ed1n+fCkiVLePjh17BaVwBpQBHvvXcjOt3jvPji0509PNU5Uqmv4C9rdyK578Ct9Gdf3df8r3glS2YOI9x8bu+gc44rw5jsF0iyX8c8P+5MXz/9AFV5g1Dc/0PBF3v9D2z/Yg4jbnkMny7WaLCjXFAB6t//fp/MzBCs1mWAgCyDzTaLu+8exTXXXHXa5IvOJssyRUVFlJXV4OdnIiEhrtNbYjz11GtYrf/AE5wAYrBa3+fNNwfw7LOPdUhWZEeSJIlffvkFi8XCuHHjvKaTcFeSGK8w74d87K6PgZkAONzX4eYvfFK/j4W3nGHr9TZw2R3wzYoTglR71Tsd7K+tocGlkOhrIF7u3L+DzAP7qCooQnGv51il9cuQ3X8ib9sqek++sTOHd1L2hmpqiw6iNwcQFJfaVEKpI11QAeqLL5Zjtd7MsUItAL3QalPZsmUL48eP76SRnZ4kSfz000by8vTodBG4XI2YTL8yc+ZgfH19WbNmDZIkMWbMmPMaaIuK8vD0lmouHpfLTX19PSEh3lMzb+vWrUydegWSFAsE43T+gZdeeo65c+/s7KF1GUkJYBg9lLIlvwIzWvzM5b6VH7bPQhx67gKUzuX0JBedRZDKa6zj0xw7kjwaQQhmQ/kODkvZvHCVi8riEtat301FhaeSxvlSVlyIKPYCjpvqVy7GWvPreRtHWyiKwqHVn5O/bTmCZiSQh85oYfDv/4o5qGPKWx11QQWo4OAA4PiXWhVkuYKAgICTrdJhZFmmtLSUmspKTD4+xMbFtatyw+HD2eTm+hEf379pWU1NGG+9tYSXX34GtzseMOJyzeGDD97m6quvan1jHWjgwMH8/PP3QPOT/AaCgkK86u7E6XQyZcpsqqv/CRydWs3mgQdGMXz4EAYMGNCZw+sSkhIg4fJLsfYdiyw/AzQCzZMnKtCZ/UivOXcP9hVZS0raGDLT95KbvQWtzZ94H/82JxPIisLS/Aa04p0E6D3TZopyMXsqP2fOM/9m5Y4H0eqG4ZAOovHREXHdfRh8z/3fcffefXG7tuJp+HHsXCRqlxEUn3jO998e5Ye3UrB9J7I7E9xhgIJbepXtXyxk5G1PdWhixwUVoObOvZmffrodq3UmEAMoCMKbRESY6N+//+lWPyNOp5PDh7NY+d0q/GvLGBAZSp1eT6afH0NnzGhzYDx8uIygoLQWy4zGAJ566l0cjjeAWUeW7uTmmycyaNBAkpOTO/ZgTuKFFx5l7Nip2GwSijIZ2I7Z/AAvv/yCV9X0W7VqFS5XIseCE0AydvudvPfeYt56Sw1Qp3I0OOlGziCzTktkn+EU7VmA7H4Fz1V/PRr9fOLGTmbRlvxzMob6sgIOr1kFuzdzqa+TIYqdzRX5ZBhDmBjX44TW7idT6bDR4AonxHDsmY4gCNQ5+rB163pc7iwcjiBAxuWcz66v32PInAfOyfE0Fx4VTa9x09i/eiKy9BwQiSC+j1b/M3H9nznn+2+P/PR1uKVH8bS8BxBAuQdb/UtYqoqa+mZ1hC4doH766SfeefllKkpLGTt1KvPuu++U7d0nTpzI/Pl/4skn+2AwDEeWCwkOdvPDD9+ck3ROu93OsmXr2LdHQsz1xyJpeOPnVQyJF7h4wAB2rV3LmOltmw7R6TTIx5Vn2bx5E56SnVOaLe2HyzWHxYs/4fHHH+2wY2nNoEGDWLt2OY8++jzp6a+TlJTEY4+9y5QpU06/8nnU0NCAopxYoV2WQ6mtvbCbH9pcLn4ozGRPRREGjZZRsSmMDItp+k40D07pdVoWby3CeNHl6AvfxFGbiCD0RpE345s4hFpDKhva0aairSzFB6ncmQNVPUk1XMuaih0Uajdyb5gv39RWkFEXQp+g01fl1goCiuI8oYRafn0DLvdE4OjdkgjKY9SVvIWjsea83EVNuP2vuIU3yd82D5fDQli3NFJGPY7e1Dkp/q1xOezA8VP3IoIQiMtp69B9ddkAtfDVV3ll/nwWHHnh9bOMDIYvXsym3btP2SrikUfu5/bbb2HDhg0EBwczYsSIVq/0XS4XDocDo9F4Rmng+/Ydoro6BqOrhsLDu9iRvguHqzs7t2wj9utn6T57GkMmTmxTokOvXjF8990hfH2Dm8ZSVpaFouiAlutLUjg1NdXtHu+ZGjBgAN9//9l529+ZGDduHJJ0B55a9nFHlkr4+Czid7+7vxNH1rmcbjdPpv9CT7uF52SZauDphlpy6qq4ofuxRntaowFBq0cQBRBAZ/JlzK2P0VCWi62+Av+IWads2yDZG1EUBZ3Rt90Xg7JLYtfBw5iNc9FRhm+9Fdk9kZ3I/Kd6Ob5CLQe1xW0KUEF6IzHmYspsewk0eJ6duhUHkvwrtCgfC2BAEH1xOe2cjxaMgigSP2AS8QMmnYe9nbmI1DQaK99Fdk3l2PP8rQhCGf4RiR26ry4ZoCwWC4/Nn88Wq5WjVfkmOJ3cVl3Nwldf5fGnT53eHBoaysyZM1v9uaIoHNi9m/xt29BLEk6DgZThw+memtqucWZlVRIUNJiM9MNs3bYT3HegwQcbQdhcs/ly6f3cVlDQpkoQiYkJDB1aS3r6KgQhDEVpZORIM19+uRZPXfajVzQOfH0/ZcaMV9s11t+68PBwnn76cf72txHYbHNRlGB8fD5g2LAIZs2adbrVf7PWlBcQZbfylSw3nWomy25SinOYEtedMOPJO8qCZ2rMPzIJ/8jWKzc4LLXUb/kec0UeIlAXHIXv4Ont6kFkrStHdkWjNYbicuYiu01AFAbGcUgpo49SxKbKEq5oQ0EVQRC4Ij6ET3L+S5V9CwghKMp+BkbksbVsA5L7D80+vRqNXsEcFNHmsV4I4gdcSvGeZ7DXX4pbuh6ELETN2/SechOipmOzdrtkgNq9ezfJWi3Hl4y92uHg+R9+gNMEqNM5tH8/dWvXMj46GoNOh9XhYOuKFeiNRhJO8uJta0wmHfX1DrJKSql1dyMAE+BCwI0VkWpxNMuXL29TgBIEgSFD+tGrVwO1tbUYjbGEho4iK2sfb789DItlHmDCx+cdxo9PY8KECWd8/L9V9913DyNHDuW99z6iru4AV1/9F6644ooL+oXig1Wl/P5IPbqjgoGxgkhGXdUpA9TpyLKbul//w1XWOvr6BSMAh+sr+Hj1J+im3on2FIVqm9PoDEADGq2BWpdEA6H4AS4aMSCzjmgy7Ydwut3o2zDTEWQw8qeesRRYqrC7y4g0+ZAY14frVy6n2nI1NtssBPEAiG+Qdtmd5yR9uivT6k0Mv/lRSvavoyLzIwx+fsT3n49vWNzpV27vvjp8i+dBeHg4BZKEBDSP11lARHT0WW1bURRy09MZFRGB4cg7PGaDgb4hIezevr1dAapv33i++y4DH18zeYIfwUolRvIQaSSPnjjFkHZPHfr5+bWoKfjii08zadJY3n9/CXa7k+vk3SHzAAALS0lEQVSuu5/f/e53F3yJlNYMGzaMYcPOsG3qb5C/wUwmAnAspVrB810a1MYA0prG8jx6NtZwsf+x5xU9zAEMrqtkW/HhNldIMPqF4BPiwlq9Abug5aCiIZAS4HtsSBRwDfAr7fmT1wgCib7HEpSCjLDl7YdYfKCar1d+Tr3eB2fQA4Qld2v7Ri8gGp2B2IsnEHvxub0Q7pIBKiUlhYv69eORbdt4RpLQAweA58xmFt9331ltW5ZlXFYr5uMqm/sYDORlZ7P+hx+QJYmIHj1I7tbtlFffCQnxjBrVSHX1Yb7UfEupKwKIxvNyYwNGfmDWrJfOaryCIDBp0iQmTfLueWuVd7okJokFxTnMlN2MANzA60CjVkefwNaf5baFZLcQeZJ3ifxlF/UHNyEc2oJs9sfQYwh+4a1XVhcEgeThY8nasBKtTwW2xhBsaIGpwLVohIVcHBTfooDtmQjwMXPP3KsZdZ3Aoi355yTZQ9U+XTJAAXzyzTfcMHs2cdu3E6XTUSTLPPvii4wbN+6stqvRaPCNiqKiro7wZkHql5070dTV0T0iAq1GQ+6qVWzMzmbklCmtJlkIgkC/fn1ITU1Blot4/vl7EYTfIQjfoCjLeOutV4mKOvvGbirVmYox+/GnPoOZfSCdQEWhXlEINpp58KLhiGd5F24KDGefAJcoMpoj02QNkpPVVUXMdEn09g+lrrGKH4sOUTNkBkFJrbcK1Jv9SZ0wA23iISrXf4yzVovbFYxB+zt8xUzuSh18VmNVeacuG6DCwsL4cd06cnNzqayspE+fPh1WRTt12DB2fv01qZJEoI8PeRUVHMjN5Q9TphDg7w9AsJ8fG3NzKS4uPm3LDqPRyOOPL+Dmm+ewbNkyNBoNs2c/S/RZTkeqVB1hSGg0A0ZGkmepw6jRdljlclNAOBUJffksZxejjD5oBIHPKwoZrtUzLjQWQRCIwES45OSVnSuQ43ud8iG7IAgYQ+K46qWP6OXMxpG3jyTXQHrkR1JYoD4n+i3qsgHqqNYqhp+NiIgIBl11FVl795JdWYk1JoahAwY0BaejInU6qsvK2txTKjExkXnz5nXoWFWqjqAVRVL8OvZdH0EQCB08jYzweDKydyHIMrZAJ1dqtC2ekQbq9ITbG7FY6jD5n35aURBFho65hIFB45E3f0dWYUmHjlvlPbp8gDpXgoODCR4zBoCKigoySk78EjS6XBi9vMCsStURXHYHOpcTRdaCAiPGpbX9GU1gDxjQAwDnzu8pKT1MgHgsAcOlyJQ6XVitoHFbT7kpWQEUUGQ3isvlKR6r+s1SA1QbhIaG4oqOJqusjKSwMERRpKymhmKjkTEJp26brVJ1dTl5wDcrSAAGjpwBg2NYvLWIEePSTrfqCWq7+7D5o1fo5Wsk3GjG4Xazsrqc4IGXMHDykNNvQIEbB8cwMMCFtP5b8jqgsrnKe6kBqg0EQWDopEnsXLeOzOxsNIAmPJxBY8eq3WNVF4STBSnhjLLm4tkTZ+S7j9/Ht7GBRlEgZuY0Hr7uJgyG06e1K7JbDU4XEDVAtZHZbGbEpEnYbDZkWfb63lEqVUc7PkgJ2jNLTBh46RDmjB9ARWUVvj4++Pkd7YJ7+vYWiksNThcSNUC1k3rHpLqQNQ9SWuOZV6gTgaMFhOR2rOeyO9TgdAFRA5RKpWqXo0GqU/evuiCoAUqlUrWbGiRU54P6dptKpVKpvJIaoFQqlUrlldQApVKpVCqvpAYolUqlUnklNUCpVCqVyiupAUqlUqlUXkkNUCqVSqXySup7UGeourqaTz/9lKL8fIaNHMm0adNO2V1XpVKdXGZDDVvKi9GKIiMjYjusH5Wq61PPqGdg27ZtTLvkEi51uUi12Xj2rbd4pWdP/vfrr2qNPpWqHT4+vJv1xTncJLuxCQIL8g9ybUpfJsemdPbQVF5AneJrJ0VRuO3aa3m1oYGPbTYWABsbGwnZt4/XXnmls4enUnUZh+qq2Vicw17ZzXPAq4rCNlnmk6w9VDtsnT08VRs5LLW4pXPTl0u9g2qn3NxcyouLuabZMhH4P7ud+z76iEcefbSzhqZSdSmbK4q4RXYT3GxZMjAFgfSqUi6NTuqsoXUJRbt/Ye//3jxhudE/jNF3vM6qf9yI7JKalvtHJjP85hfatO28bT+Qs2kpiiKTNHQWiUOmA9BYUUD6589hCghl4DUL0Gj15Kf/QMqIKzvmoI6jBqh20mg0uBTlhMYAEqjPoFSqdhAFAedJlksCiII6uXM6UX1GEd6jZZPHjJUfoNHpaSjPRWvwYdRdrzb9TBDb9jst2b+eQ798RNr0ufiFJbBz6UuYgyII7z6Yor2rie4ziprCg1Tn7yMgMhmtwYyo1XXosR2l/hW0U3x8PN26d+cdQWha5gSeN5m45rbbOm9gKlUXMzIijg9EDUXNlu0GVikKg0MjO2tYXYao0aEz+jT947TWUZGZTsqoa6gtPEhQXK8WP9fq29YqKHfLt0T3HUdUr5H4hsbSbdTV5G37AQDJ1oApMBKTfyiStYHC3b8Q03fcuTvGc7bl37D3P/uMF0JDudTPj3l6Pb19fPAbM4Y/33VXZw9NpeoyEn0DmJ6YSpoocosoco2oYbQockfqIPx1Z95r6kKVuWYJ8QOnYPAJoLboIPWl2fz86i2sfOVGdi17Fae1/rTbUBSFhvJcQpP7NS0LiOlBXclhALQGM5K9Aae9AVGrw+20ozcHnLNjUuekzkBqaiqHCgr45ptvKCoq4uPhwxk6dChCs7sqlUp1epcn9GRYeCzbqkoJEUX+GRpNgF4NTu1lrS2jInsHvSbdCoClppSIHkOIHzgVydbI3v+9wYEV73Px5f93yu24JTuK7MYcGNG0TGfwwe20I9ktRPQczvbPn0Fn8sdeX0lU71Hn9LjUAHWGDAYDV199dWcPQ6Xq8iJMPkxT08rPSt7W74nsNbLpbmbELX9v+pkpIIyel9xE+mfP4JYcaE5xdyqIGgBErb5pmUbn+Xe3ZCcotidj576DqNFy+Nf/UJ2/j/TPniF5+Gzi+k/q8ONSp/hUKpWqC5PdEiX71xHdZ0yrnzH4BaPIbhyW2lNuS6PVo9EZcDb7nOSwAp5nXgBavYmq3L2EJvcna93npE64mcx1n539gZyEGqBUKpWqC6vI2oFGqycorhcAttpy1v/7r8huV9NnagszEDU6jH4hp91eYEwPagozmv67vjQbjc6AzuTbtKw6bw8hiX1x2uoJ6zYQydrQgUd0jBqgVCqVqgurOLyNkMSLmp6BmwLDURQ3+3/6F/WlOZQe2MCh1Z8QP3AKosbzVMflsCK7pZNuL6r3aPK2/Q9bXQVul5PsjUsJTe6PcCT1v6E8D9+wBAB0Rl9qCg60CF4dSQ1QKpVK1YVV5u4iKL53i2X9r3gAa00pWz55lMx1n5E4eDrdx17X9PP1//4rBTtXnnR70X3HEhzfh3Xv3cPqf96OrbaUnuNvOLa/nJ1E9R4BQOLQmexc+iKJQy8/B0emJkmoVCpVlzZu7jsnLPMJjmLIdY+3us7YP59YgeIoQRC5+PL/o740G0djDUFxvdEajr1DldQsGCUPm0XysFlnNvA2UAOUSqVSqU7gH5nc2UNQp/hUKpVK5Z3UAKVSqVQqr6QGKJVKpVJ5JTVAqVQqlcorqQFKpVKpVF5JDVAqlUql8kpqgFKpVCqVVxIU5fjesKf4sCBUAHnnbjgqlddLUBQl7Gw2oH6PVKq2fY/aFaBUKpVKpTpf1Ck+lUqlUnklNUCpVCqVyiupAUqlUqlUXkkNUCqVSqXySmqAUqlUKpVXUgOUSqVSqbySGqBUKpVK5ZXUAKVSqVQqr6QGKJVKpVJ5pf8HwLrvLvz2RyQAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x648 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#创建画布，宽高比为6*9\n",
    "figure = plt.figure(figsize=(6, 9))\n",
    "#设置用来安排图像显示位置的全局变量i\n",
    "i = 1\n",
    "#开始迭代数据，对datasets中的数据进行for循环\n",
    "for ds_index, ds in enumerate(datasets):\n",
    "    X, y = ds\n",
    "    # 对数据进行标准化处理\n",
    "    X = StandardScaler().fit_transform(X)\n",
    "    # print(X)\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, random_state=42)\n",
    "    # 找出数据集中两个特征的最大值和最小值，让最大值+0.5，最小值-0.5，创造一个比两个特征的区间本身更大\n",
    "    # 一点的区间\n",
    "    x1_min, x1_max = X[:,0].min() - 0.5, X[:,0].max() + 0.5\n",
    "    x2_min, x2_max = X[:,1].min() - 0.5, X[:,1].max() + 0.5\n",
    "    #用特征向量生成网格数据，网格数据，其实就相当于坐标轴上无数个点\n",
    "    #函数np.arange在给定的两个数之间返回均匀间隔的值，0.2为步长\n",
    "    #函数meshgrid用以生成网格数据，能够将两个一维数组生成两个二维矩阵。\n",
    "    #如果第一个数组是narray，维度是n，第二个参数是marray，维度是m。那么生成的第一个二维数组是以\n",
    "    #narray为行，m行的矩阵，而第二个二维数组是以marray的转置为列，n列的矩阵\n",
    "    #生成的网格数据，是用来绘制决策边界的，因为绘制决策边界的函数contourf要求输入的两个特征都必须是二\n",
    "    #维的\n",
    "    array1,array2 = np.meshgrid(np.arange(x1_min, x1_max, 0.2),\n",
    "                                np.arange(x2_min, x2_max, 0.2))\n",
    "    \n",
    "    #接下来生成彩色画布\n",
    "    #用ListedColormap为画布创建颜色，#FF0000正红，#0000FF正蓝\n",
    "    cm = plt.cm.RdBu\n",
    "    cm_bright = ListedColormap(['#FF0000', '#0000FF'])\n",
    "    \n",
    "    #在画布上加上一个子图，数据为len(datasets)行，2列，放在位置i上\n",
    "    ax = plt.subplot(len(datasets), 2, i)\n",
    "    \n",
    "    #到这里为止，已经生成了0~1之间的坐标系3个了，接下来为我们的坐标系放上标题\n",
    "    #我们有三个坐标系，但我们只需要在第一个坐标系上有标题，因此设定if ds_index==0这个条件\n",
    "    if ds_index == 0:\n",
    "        ax.set_title(\"Input data\")\n",
    "    #将数据集的分布放到我们的坐标系上\n",
    "    #先放训练集\n",
    "    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train,\n",
    "                cmap=cm_bright,edgecolors='k')\n",
    "    \n",
    "    #放测试集\n",
    "    ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test,\n",
    "                cmap=cm_bright, alpha=0.3,edgecolors='k')\n",
    "    #为图设置坐标轴的最大值和最小值，并设定没有坐标轴\n",
    "    ax.set_xlim(array1.min(), array1.max())\n",
    "    ax.set_ylim(array2.min(), array2.max())\n",
    "    ax.set_xticks(())\n",
    "    ax.set_yticks(())\n",
    "    #每次循环之后，改变i的取值让图每次位列不同的位置\n",
    "    i += 1\n",
    "    \n",
    "    #至此为止，数据集本身的图像已经布置完毕，运行以上的代码，可以看见三个已经处理好的数据集\n",
    "    \n",
    "    #############################从这里开始是决策树模型##########################\n",
    "    #迭代决策树，首先用subplot增加子图，subplot(行，列，索引)这样的结构，并使用索引i定义图的位置\n",
    "    #在这里，len(datasets)其实就是3，2是两列\n",
    "    #在函数最开始，我们定义了i=1，并且在上边建立数据集的图像的时候，已经让i+1,所以i在每次循环中的取值\n",
    "    #是2，4，6\n",
    "    ax = plt.subplot(len(datasets),2,i)\n",
    "    #决策树的建模过程：实例化 → fit训练 → score接口得到预测的准确率\n",
    "    clf = DecisionTreeClassifier(max_depth=5)\n",
    "    clf.fit(X_train, y_train)\n",
    "    score = clf.score(X_test, y_test)\n",
    "    \n",
    "    #在这里，我们先将两个网格数据降维降维成一维数组，再将两个数组链接变成含有两个特征的数据，再带入决策\n",
    "    #树模型，生成的Z包含数据的索引和每个样本点对应的类概率，再切片，切出类概率\n",
    "    Z = clf.predict_proba(np.c_[array1.ravel(),array2.ravel()])[:, 1]\n",
    "\n",
    "    #将返回的类概率作为数据，放到contourf里面绘制去绘制轮廓\n",
    "    Z = Z.reshape(array1.shape)\n",
    "    ax.contourf(array1, array2, Z, cmap=cm, alpha=.8)\n",
    "    #将数据集的分布放到我们的坐标系上\n",
    "    # 将训练集放到图中去\n",
    "    ax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright,\n",
    "    edgecolors='k')\n",
    "    # 将测试集放到图中去\n",
    "    ax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright,\n",
    "    edgecolors='k', alpha=0.6)\n",
    "    #为图设置坐标轴的最大值和最小值\n",
    "    ax.set_xlim(array1.min(), array1.max())\n",
    "    ax.set_ylim(array2.min(), array2.max())\n",
    "    #设定坐标轴不显示标尺也不显示数字\n",
    "    ax.set_xticks(())\n",
    "    ax.set_yticks(())\n",
    "    #我们有三个坐标系，但我们只需要在第一个坐标系上有标题，因此设定if ds_index==0这个条件\n",
    "    if ds_index == 0:\n",
    "        ax.set_title(\"Decision Tree\")\n",
    "    #写在右下角的数字\n",
    "    ax.text(array1.max() - .3, array2.min() + .3, ('{:.1f}%'.format(score*100)),\n",
    "    size=15, horizontalalignment='right')\n",
    "    #让i继续加一\n",
    "    i += 1\n",
    "    \n",
    "plt.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 206,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T08:25:30.974638Z",
     "start_time": "2019-12-05T08:25:30.965713Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[array([[  1,  10, 100],\n",
       "        [  1,  10, 100]]), array([[ 2,  2,  2],\n",
       "        [20, 20, 20]])]"
      ]
     },
     "execution_count": 206,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.meshgrid([1,10,100],[2,20])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-12-05T14:10:18.586108Z",
     "start_time": "2019-12-05T14:10:18.576732Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1, 4],\n",
       "       [2, 5],\n",
       "       [3, 6]])"
      ]
     },
     "execution_count": 218,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.c_[np.array([1,2,3]), np.array([4,5,6])]"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}