Tutorial V2 - Sentiment.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Movie Reviews Dataset\n",
    "* Input patterns: movie reviews in natural language.\n",
    "* Labels: sentiment analysis of the text\n",
    "    * Classes: *negative, positive*\n",
    "* Ground truth: What the person thought when writing the review\n",
    "* Annotations: re-collected from Amazon Mechanical Turk (AMT).\n",
    "    * Number of annotators (T) = 203  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import keras, time, sys, os, gc\n",
    "np.set_printoptions(precision=4, suppress=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#GLOBAL Variables\n",
    "BATCH_SIZE = 128 #rodrigues = BATCH_SIZE = 64\n",
    "EPOCHS_BASE = 50\n",
    "OPT = 'adam' #optimizer for neural network \n",
    "TOL = 1e-2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "folder = \"./data/Sentiment/\""
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load Data and preprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Num. train texts: 4999\n",
      "Num. test texts:  3999\n"
     ]
    }
   ],
   "source": [
    "from codeE.utils import read_texts\n",
    "texts_train = read_texts(folder+\"Sentiment_text_train.txt\")\n",
    "texts_test  = read_texts(folder+\"Sentiment_text_test.txt\")\n",
    "\n",
    "Z_train = np.loadtxt(folder+\"Sentiment_Z_train.txt\", dtype='int')\n",
    "Z_test  = np.loadtxt(folder+\"Sentiment_Z_test.txt\", dtype='int')\n",
    "\n",
    "print(\"Num. train texts: %d\" % len(texts_train))\n",
    "print(\"Num. test texts:  %d\" % len(texts_test))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 14096 unique tokens.\n"
     ]
    }
   ],
   "source": [
    "from keras import preprocessing\n",
    "MAX_NB_WORDS = 14000\n",
    "tokenizer = preprocessing.text.Tokenizer(num_words=MAX_NB_WORDS)\n",
    "tokenizer.fit_on_texts(texts_train+texts_test)\n",
    "sequences_train = tokenizer.texts_to_sequences(texts_train)\n",
    "sequences_test = tokenizer.texts_to_sequences(texts_test)\n",
    "print('Found %s unique tokens.' % len(tokenizer.word_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Real max:  51\n",
      "Used max:  50\n",
      "Shape of train tensor: (4999, 50)\n"
     ]
    }
   ],
   "source": [
    "lens = list(map(len,sequences_train))\n",
    "max_L = np.max(lens)\n",
    "print(\"Real max: \",max_L)\n",
    "max_L = 50 \n",
    "print(\"Used max: \",max_L)\n",
    "\n",
    "X_train = preprocessing.sequence.pad_sequences(sequences_train, maxlen=max_L,dtype='int32', value=0,padding='pre')\n",
    "X_test = preprocessing.sequence.pad_sequences(sequences_test, maxlen=max_L,dtype='int32', value=0,padding='pre')\n",
    "print('Shape of train tensor:', X_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Preparing embedding matrix.\n",
      "Words found on glove:  12935\n"
     ]
    }
   ],
   "source": [
    "EMBEDDING_DIM = 300\n",
    "GLOVE_FILE = \"../GLOVE/glove.6B.%dd.txt\"%(EMBEDDING_DIM)\n",
    "\n",
    "embeddings_index = {}\n",
    "with open(GLOVE_FILE) as file:\n",
    "    for line in file:\n",
    "        values = line.split()\n",
    "        coefs = np.asarray(values[1:], dtype='float32')\n",
    "        embeddings_index[values[0]] = coefs\n",
    "print('Preparing embedding matrix.')\n",
    "sorted_x = sorted(tokenizer.word_counts.items(), key=lambda kv: kv[1], reverse=True)\n",
    "vocab = {value[0]:tokenizer.word_index[value[0]] for i, value in enumerate(sorted_x) if i < MAX_NB_WORDS}\n",
    "embedding_matrix = np.zeros((len(vocab)+1, EMBEDDING_DIM))\n",
    "v=0\n",
    "for word, i in vocab.items():\n",
    "    embedding_vector = embeddings_index.get(word)\n",
    "    if embedding_vector is not None:\n",
    "        embedding_matrix[i] = embedding_vector # words not found in embedding index will be all-zeros.\n",
    "        v+=1\n",
    "del embeddings_index, sorted_x, tokenizer\n",
    "gc.collect()\n",
    "print(\"Words found on glove: \",v)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Load annotations "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading AMT data...\n",
      "Shape (data,annotators):  (4999, 203)\n",
      "Classes:  2\n"
     ]
    }
   ],
   "source": [
    "print(\"Loading AMT data...\")\n",
    "y_obs = np.loadtxt(folder+\"answers.txt\",dtype='int16') #not annotation symbol ==-1\n",
    "\n",
    "N,T = y_obs.shape\n",
    "print(\"Shape (data,annotators): \",(N,T))\n",
    "print(\"Classes: \", np.max(y_obs)+1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Upper Bound Model -- ideal"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#### Delta Convergence criteria\n",
    "from codeE.learning_models import default_RNN_text, Clonable_Model\n",
    "from codeE.utils import EarlyStopRelative\n",
    "ourCallback = EarlyStopRelative(monitor='loss',patience=1,min_delta=TOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n",
      "WARNING:tensorflow:From /home/casapanshop/anaconda2/envs/newpy3_tf1/lib/python3.6/site-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
      "Trained IDeal Model, Epochs to converge = 50\n"
     ]
    }
   ],
   "source": [
    "Z_train_onehot = keras.utils.to_categorical(Z_train)\n",
    "\n",
    "model_UB = default_RNN_text(max_L, Z_train_onehot.shape[1], embed_M=embedding_matrix)\n",
    "model_UB.compile(loss='categorical_crossentropy',optimizer=OPT)\n",
    "hist = model_UB.fit(X_train,Z_train_onehot,epochs=EPOCHS_BASE,batch_size=BATCH_SIZE,verbose=0,callbacks=[ourCallback])\n",
    "print(\"Trained IDeal Model, Epochs to converge =\",len(hist.epoch))\n",
    "clone_UB = Clonable_Model(model_UB)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy of model = 1.0\n",
      "Test accuracy of model = 0.7214303575893973\n"
     ]
    }
   ],
   "source": [
    "from codeE.evaluation import accuracy_model\n",
    "print(\"Train accuracy of model =\", accuracy_model(model_UB, X_train, Z_train))\n",
    "print(\"Test accuracy of model =\", accuracy_model(model_UB, X_test, Z_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Generate Matrix representation of both scenarios"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Individual representation shape (N,T,K)=  (4999, 203, 2)\n",
      "Global representation shape (N,K)=  (4999, 2)\n"
     ]
    }
   ],
   "source": [
    "from codeE.representation import set_representation\n",
    "\n",
    "y_obs_categorical = set_representation(y_obs,'onehot') \n",
    "print(\"Individual representation shape (N,T,K)= \",y_obs_categorical.shape)\n",
    "\n",
    "r_obs = set_representation(y_obs,\"global\")\n",
    "print(\"Global representation shape (N,K)= \",r_obs.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Rodrigues 18' (Raykar without EM)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 50, 300)           4200300   \n",
      "_________________________________________________________________\n",
      "cu_dnngru_1 (CuDNNGRU)       (None, 50, 128)           165120    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 50, 128)           0         \n",
      "_________________________________________________________________\n",
      "cu_dnngru_2 (CuDNNGRU)       (None, 64)                37248     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 2)                 130       \n",
      "=================================================================\n",
      "Total params: 4,402,798\n",
      "Trainable params: 202,498\n",
      "Non-trainable params: 4,200,300\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model_R = clone_UB.get_model()\n",
    "model_R.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 183us/step - loss: 3.0843\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9889\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9190\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8742\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8273\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7489\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6997\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.6528\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6021\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.5673\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.5194\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.5114\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.1174\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0127\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9235\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.8705\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.8112\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7763\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.7216\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.6709\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.6295\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.5840\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.5222\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.5074\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 166us/step - loss: 3.0710\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9744\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.8975\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8177\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7582\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.6986\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6798\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.1509\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0189\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9422\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9717\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8640\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.7815\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.7208\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.7110\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.1002\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.9963\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.9487\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8880\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8268\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.7898\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.7563\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.6881\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.6371\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.5917\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.6286\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.5509\n",
      "Epoch 13/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.5264\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 171us/step - loss: 3.1698\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0350\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9895\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9461\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.8565\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.8220\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 86us/step - loss: 2.7620\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7121\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6492\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6106\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.5456\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.5055\n",
      "Epoch 13/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.4878\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 173us/step - loss: 3.1481\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 3.0352\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9242\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.8914\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8747\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 167us/step - loss: 3.1001\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9791\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9236\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8591\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.7940\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.7374\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.6926\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6378\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6149\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 170us/step - loss: 3.1390\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 3.0396\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9614\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9240\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.8514\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.7990\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.7395\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6878\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7482\n",
      "Epoch 10/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.6147\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.5806\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.5361\n",
      "Epoch 13/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.4905\n",
      "Epoch 14/50\n",
      "4999/4999 [==============================] - 0s 86us/step - loss: 2.4602\n",
      "Epoch 15/50\n",
      "4999/4999 [==============================] - 0s 86us/step - loss: 2.4300\n",
      "Epoch 16/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.4065\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 170us/step - loss: 3.1693\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0128\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9510\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8863\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8616\n",
      "Finished training\n",
      "10 runs over Rodrigues 18', Epochs to converge=  10.0\n"
     ]
    }
   ],
   "source": [
    "from codeE.methods import ModelInf_BP as Rodrigues18\n",
    "#Ro_model = Rodrigues18(init_conf=\"default\",\n",
    "Ro_model = Rodrigues18(init_Z=\"softmv\", n_init_Z= 3, init_conf=\"model\",\n",
    "                       prior_lamb = 0) #prior on lambda KL restriction on MV\n",
    "args = {'batch_size':BATCH_SIZE, 'optimizer':OPT} \n",
    "Ro_model.set_model(model_R, **args)\n",
    "\n",
    "logL_hists, i_r = Ro_model.fit(X_train, y_obs_categorical, runs=10, max_iter=EPOCHS_BASE, tolerance=TOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(15,7))\n",
    "for a, value in enumerate(logL_hists): \n",
    "    if a != i_r:\n",
    "        plt.plot(range(len(value)),value,'.-')\n",
    "plt.plot(range(len(logL_hists[i_r])),logL_hists[i_r],'o-',markersize=10,label=\"Selected run\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy of model = 0.885377075415083\n",
      "Test accuracy of model = 0.75743935983996\n"
     ]
    }
   ],
   "source": [
    "raykar_fx = Ro_model.get_basemodel()\n",
    "\n",
    "from codeE.evaluation import accuracy_model\n",
    "print(\"Train accuracy of model =\", accuracy_model(raykar_fx, X_train, Z_train))\n",
    "print(\"Test accuracy of model =\", accuracy_model(raykar_fx, X_test, Z_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Global Behavior - Label Noise\n",
    "Based on Noise Channel of Bekker and Goldberg"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 50, 300)           4200300   \n",
      "_________________________________________________________________\n",
      "cu_dnngru_1 (CuDNNGRU)       (None, 50, 128)           165120    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 50, 128)           0         \n",
      "_________________________________________________________________\n",
      "cu_dnngru_2 (CuDNNGRU)       (None, 64)                37248     \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 2)                 130       \n",
      "=================================================================\n",
      "Total params: 4,402,798\n",
      "Trainable params: 202,498\n",
      "Non-trainable params: 4,200,300\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "aux_model = clone_UB.get_model()\n",
    "aux_model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 197us/step - loss: 3.2668\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.1630\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 3.1130\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 86us/step - loss: 3.0587\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.9722\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 85us/step - loss: 2.9350\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9004\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8169\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8202\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.3109\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 3.2080\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 78us/step - loss: 3.1526\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 78us/step - loss: 3.1086\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 77us/step - loss: 3.0387\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.9753\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.9373\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.8877\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8297\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 78us/step - loss: 2.7797\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.7377\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.6890\n",
      "Epoch 13/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.6671\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 177us/step - loss: 3.2102\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 3.1124\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.0675\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.9986\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.9588\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9903\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.8464\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8084\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7413\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.7054\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.6550\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.6780\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 177us/step - loss: 3.2573\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.1577\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0801\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.0979\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 168us/step - loss: 3.2558\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 3.1546\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.0878\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 3.0235\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9846\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 78us/step - loss: 2.9331\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 78us/step - loss: 2.9098\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 172us/step - loss: 3.2682\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.1409\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0797\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.0370\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9666\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.9381\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 171us/step - loss: 3.2726\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.1745\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0912\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 3.0402\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9890\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9180\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.8461\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.8162\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.7617\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.7134\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.6653\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.6513\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 167us/step - loss: 3.2479\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.1489\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.1036\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 3.0552\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.9689\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 83us/step - loss: 2.9394\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.2308\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 3.1286\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 3.0533\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 3.0195\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.9550\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.8901\n",
      "Epoch 7/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.8513\n",
      "Epoch 8/50\n",
      "4999/4999 [==============================] - 0s 81us/step - loss: 2.7953\n",
      "Epoch 9/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.7381\n",
      "Epoch 10/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.7025\n",
      "Epoch 11/50\n",
      "4999/4999 [==============================] - 0s 79us/step - loss: 2.6597\n",
      "Epoch 12/50\n",
      "4999/4999 [==============================] - 0s 80us/step - loss: 2.6618\n",
      "Finished training\n",
      "Initializing...\n",
      "Pre-train network base_model_z on 3 epochs... Done!\n",
      "Epoch 1/50\n",
      "4999/4999 [==============================] - 1s 169us/step - loss: 3.2476\n",
      "Epoch 2/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.1476\n",
      "Epoch 3/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0979\n",
      "Epoch 4/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 3.0403\n",
      "Epoch 5/50\n",
      "4999/4999 [==============================] - 0s 82us/step - loss: 2.9733\n",
      "Epoch 6/50\n",
      "4999/4999 [==============================] - 0s 84us/step - loss: 2.9627\n",
      "Finished training\n",
      "10 runs, Epochs to converge=  8.7\n"
     ]
    }
   ],
   "source": [
    "from codeE.methods import ModelInf_BP_G as G_Noise_BP\n",
    "#GNoise_model = G_Noise_BP(init_conf = \"default\") \n",
    "GNoise_model = G_Noise_BP(init_Z='softmv', n_init_Z=3, init_conf = \"model\") \n",
    "args = {'batch_size':BATCH_SIZE, 'optimizer':OPT}\n",
    "GNoise_model.set_model(aux_model, **args)\n",
    "\n",
    "loss_hists,i_r = GNoise_model.fit(X_train, r_obs, runs =10, max_iter=EPOCHS_BASE,tolerance=TOL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train accuracy of model = 0.8709741948389678\n",
      "Test accuracy of model = 0.7549387346836709\n"
     ]
    }
   ],
   "source": [
    "G_fx = GNoise_model.get_basemodel()\n",
    "\n",
    "from codeE.evaluation import accuracy_model\n",
    "print(\"Train accuracy of model =\", accuracy_model(G_fx, X_train, Z_train))\n",
    "print(\"Test accuracy of model =\", accuracy_model(G_fx, X_test, Z_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Confusion Matrices Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "from codeE.utils import generate_confusionM\n",
    "from codeE.utils import plot_confusion_matrix, compare_conf_ma, compare_set_conf_ma\n",
    "from codeE.evaluation import Individual_D, D_JS, D_NormF"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Real Confusion matrices\n",
    "Generate the individual and global confusion matrices based on the sampled annotations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Individual matrix shape (T,K,K)=  (203, 2, 2)\n",
      "Global matrix shape (K,K)=  (2, 2)\n"
     ]
    }
   ],
   "source": [
    "from codeE.utils import generate_Individual_conf, generate_Global_conf\n",
    "\n",
    "conf_matrix_I = generate_Individual_conf(Z_train, y_obs_categorical)\n",
    "print(\"Individual matrix shape (T,K,K)= \",conf_matrix_I.shape)\n",
    "\n",
    "confe_matrix_G = generate_Global_conf(Z_train, r_obs)\n",
    "print(\"Global matrix shape (K,K)= \",confe_matrix_G.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Global scenario"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Global - Label Noise (with predictive model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Global D (based on Jensen Shannon) = 0.0021100736181346226\n",
      "Global D (based on normalized Frobenius) = 0.04427885264158249\n"
     ]
    }
   ],
   "source": [
    "prob_Yz = GNoise_model.get_global_confusionM()\n",
    "compare_conf_ma(prob_Yz, true_conf_mat = confe_matrix_G, text=True)\n",
    "\n",
    "print(\"Global D (based on Jensen Shannon) =\",D_JS(confe_matrix_G, prob_Yz))\n",
    "print(\"Global D (based on normalized Frobenius) =\",D_NormF(confe_matrix_G, prob_Yz))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    ">### Individual scenario\n",
    "#### Rodrigues"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot 5 random matrices from the set\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 1.2466980225295536\n",
      "D (based on normalized Frobenius) = 0.33042216300964355\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.6246232737206822\n",
      "D (based on normalized Frobenius) = 0.8145655393600464\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAADpCAYAAABxwaO9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAWiklEQVR4nO3de3RU5bnH8d9DwqU0JEBBMCRecCgVKsRjuPQUgYoFxBpSjyClFVYRscil2irWsipY76e251hIq1YUbBsj4iVigeDlGBQvETWKpGoolyZBQARilHt4zx8ZMQmJSSQz887k+1mLtWa/+509z97h8Tf7QjTnnAAA8E2rSBcAAEBdCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoADHLzM41s/cjXUddzGy4mZVGug6fEVCeMbNPq/05amb7qy3/ONL1AeFiZltq/f3/1MwWNvAeZ2aBz5edcy8653qHqL7FZnZLKLaNKvGRLgA1OecSPn9tZlskTXXOPVvffDOLd84dCUdtQARc9GV//xHbOIOKMmZ2i5k9YmYPm1mFpJ+Y2d/MbH61OecHw+3z5RQze8LMPjKzzWY2IwKlA83CzAJmlm9m5Wa2y8weCY6vCU55O3i2dWnty2jBs7LrzOwdM/vMzBaZWTczW2lmFWb2rJl1qjb/UTPbHvysNWbWNzg+TdKPJc0Jftby4HiymT1WrddmV9vW14JnXXvMrEjSgNAfrehGQEWnH0rKlpQk6ZEvm2hmcZKelvS6pB6Svi/pOjMbEeoigRC5WdJqSZ0kpUhaIEnOuaHB9f2dcwnOufp6479U1QfflHSRpJWSfi2pi6r+mzi72tyVknpJOknSm5L+Hvys+4Kv/zv4WReZWStJyyW9rapeGyHpajMbFdzWPElnBP+MkjT5BI5Bi0BARaeXnHPLnXNHnXP7G5g7WFKic+4259wh59xGSYskTQh9mcAJe9LM9lb7c4Wkw5JOlZTsnDvgnHupidtc4Jzb4Zwrk/SipNecc2855w5KekLS2Z9PdM494JyrCK6bL6m/mSXVs90Bkro6534b7LVNkv6iL3ptvKRbnXO7nXMlkv7YxLpbHO5BRaeSJsw9VdIpZra32licpBeatSIgNDJr34MKXk67WVKBme2R9Hvn3ANN2OaOaq/317GcEPycOEm3Shonqauko8E5XSSV17HdUyUl19FrLwZfJ6tm725tQs0tEgEVnWr/CvrPJLWvtty92usSScXOuTNDXhUQBs657ZKukCQzGyLpWTNbE7w60JwmShor6XxJW1R1SX2PJPu8lFrzSyRtds71qmd7H0pKlbQhuHxKcxYbi7jEFxsKJV1oZp3M7GTVvIb+iqRDZvZLM2tnZnFmdpaZnROZUoETY2bjzCwluLhHVUFRGVzeIalnM31UB0kHJX2sqi+At9VaX/uzCiR9YmbXBx+IiDOzb5vZ5w9DLJV0Q7BPUyTNaqY6YxYBFRsWS/qnqi4ZrJKU8/mK4CPoYyQNVNW3wF2S7pWUGO4iga9gea1/B/WEqu71vGZmn0p6StLPnXObg/PnS1oSvF81/gQ/+yFV9VSZpCJJr9Zav0hSn+BnPemcq1TVQxdpkjarqtfuV9WZlyTdFNzeZlU95PHXE6wv5hn/w0IAgI84gwIAeImAAgB4iYACAHiJgAIAeCkk/w6qvcW5jmodik2jHsnnNNeTtWisLVu2adeuvdbwzK+GPgo/+igy3njjn7ucc11rj4ckoDqqta7UqaHYNOoxbx1PrIZbevplId0+fRR+9FFkmKXX+Vs1uMQHAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8BIBBQDwEgEFAPASAQUA8FKLCKiMRbfp2h0va/r65fXOGX33XM0qXq2fvf2Uup/d59h4/0mZmvlBnmZ+kKf+kzLDUW7MWLXqZfXufbECgUzdccfi49YfPHhIl156gwKBTA0aNFlbtmw7tu722x9UIJCp3r0vVl7eK2GsGl+moV76Ru+emvJyjuYeWK/v/HJKjXWDZk/S9PXLNf3dpzXo55PDUW7McM5p9uzfKRDIVL9+E/Tmm+/VOW/u3Cylpl6ohIRza4xv3fqhRoyYrn79Jmj48GkqLd0RjrJPWIsIqMLFj+tvo6fWuz5wwVB17nWaFvQaqeXTfqML/zxfktSuU5KGzZup+weN1/0Dx2nYvJlq1zExTFVHt8rKSs2YcadWrvyjiooe1cMP56moaFONOYsW5apTpw7auPFJXXPNRF1//QJJUlHRJuXkrNaGDUu1atUCXXXVHaqsrIzEbqCWhnpp/+69WjX7Vr1y16Ia41379tJ/XDFOfxk4Tvf0H6tv/mC4OgdODXW5MWPlyrUqLi5RcfETuu++uZo+/fY651100VAVFCw5bvzaa/9XkyZdqHfeydGNN16hG25YGOqSm0WLCKh/v7hO+3eX17v+W2NH6J2HnpQklb32ttp1TFRC964KjBqiTc+s1YE95Tqw9xNtematAqPPrXc7+EJBwQYFAqnq2TNFbdq01oQJI5Wbm19jTm5uviZP/oEk6ZJLRui55wrknFNubr4mTBiptm3b6PTTeygQSFVBwYZI7AZqaaiX9n20W9vWrVfl4SM1xrueeYZKX31bR/YfkKus1Nb81/WtH34/1OXGjNzcfE2aNEZmpsGDz9LevRX68MNdx80bPPgsnXxyl+PGi4o2a8SIAZKk730vXbm5a0Jec3NoEQHVkA49uqm8ZPux5U9Kt6tDj251jO9Qhx7dIlFi1Ckr26nU1C+OVUrKSSor21nvnPj4eCUlJejjj8sb9V5El53vfqBTh6bra507Kv5r7RQYM1RJqd0jXVbUKCv7SKnVjldKSrcm9UT//r302GPPS5KeeOL/VFHxmT7+eG+z19ncGhVQZjbazN43s41m9qtQFxV2ZsePOSerZxwNq+sw1T6edc9p3HujVcz3Uj12vbdJa++8X5c984B+sup+7Xj7fR09wmXbxnJ1NEVTeuKuu65Wfv6bOvvsicrPf1M9epyk+Pj45iwxJBoMKDOLk5Ql6QJJfST9yMz6fPm7oktF6fYa3+YSU7qrYttOfXLceDdVbOObfGOkpJykkpIvbsSWlu5UcnLXeuccOXJE5eWfqnPnpEa9Nxq1hF76Mm89sEz3nXOxFg/7ifbv3quPi7dGuiSvZWUtVVraRKWlTVRycleVVLuaU1q6o0k9kZzcVY8//ju99Va2br31KklSUlJCs9fc3BpzBjVQ0kbn3Cbn3CFJOZLGhras8Hr/qefVL/iEXo9B/XWwvEKfbv9IG/NeUs+RQ9SuY6LadUxUz5FDtDHvpQhXGx0GDOij4uISbd5cpkOHDisnZ7UyMobWmJORMVRLljwtSVq27Dmdd94AmZkyMoYqJ2e1Dh48pM2by1RcXKKBA/tGYjeaW8z30pdp37WzJCkx9WSdefFIvfvw0xGuyG8zZoxXYWG2CguzlZk5XA89tELOOb366nolJSXUea+pPrt27dXRo0clVT0hO2VKRqjKblaNOcfrIamk2nKppEG1J5nZNEnTJCmpUZsNn4uzf6/Thg9U+y6ddE1Jvl6Yt0CtWlfV+Ma9OSpeka9eY4Zp1sZndHjffuX+9NeSpAN7yrXm5j/piteXSZLW/DZLB/bUf4MYX4iPj9fChddp1KhZqqys1JQpGerb9wzdeOM9Sk8/UxkZw3T55WN12WU3KhDIVOfOicrJuU2S1LfvGRo//nz16TNO8fFxysqao7i4uAjvUbNosJd87iOp4V76ercumrbuMbVNTJA7elSDr56srD5jdKjiM41/bIHaf6OjKg8f0YoZN+nA3k8ivDfRY8yY72rFirUKBDLVvn07PfjgvGPr0tImqrAwW5I0Z87dys7O0759B5SSMkZTp47V/PlX6oUX1umGG7JkZho69GxlZV0fqV1pEqvr2maNCWbjJI1yzk0NLl8maaBzblZ970m2du5K8QhpOM1z2ZEuocVJT79M69YVNfpGQFN7iT4KP/ooMszS33DOpdceb8wlvlJJqdWWUyRtq2cugPrRS0ATNCagXpfUy8xON7M2kiZIeiq0ZQExiV4CmqDBi9zOuSNmNlNSnqQ4SQ845/hXk0AT0UtA0zTqLqxzboWkFSGuBYh59BLQePwmCQCAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAlwgoAICXCCgAgJcIKACAl+JDsdHkc3pq3rq/hmLTqMdNNjHSJbQ427Q1pNunj8KPPvILZ1AAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvtZiAWrXqZfXufbECgUzdccfi49YfPHhIl156gwKBTA0aNFlbtmw7tu722x9UIJCp3r0vVl7eK2GsOnplLLpN1+54WdPXL693zui752pW8Wr97O2n1P3sPsfG+0/K1MwP8jTzgzz1n5QZjnLRSM45zZ79OwUCmerXb4LefPO9OufNnZul1NQLlZBwbo3xe+5ZprPOulRpaRM1ZMjlKiraFI6yo1pDvfSN3j015eUczT2wXt/55ZRj43Ft22jqa4/qysJcTX/3aQ2fPytcJTebFhFQlZWVmjHjTq1c+UcVFT2qhx/OO64xFi3KVadOHbRx45O65pqJuv76BZKkoqJNyslZrQ0blmrVqgW66qo7VFlZGYndiCqFix/X30ZPrXd94IKh6tzrNC3oNVLLp/1GF/55viSpXackDZs3U/cPGq/7B47TsHkz1a5jYpiqRkNWrlyr4uISFRc/ofvum6vp02+vc95FFw1VQcGS48YnThyt9esfUWFhtubMmaRf/OJ/Ql1y1Guol/bv3qtVs2/VK3ctqjFeefCQlpw3WfemjdW9aZk6Y/S56jGof6jLbVYtIqAKCjYoEEhVz54patOmtSZMGKnc3Pwac3Jz8zV58g8kSZdcMkLPPVcg55xyc/M1YcJItW3bRqef3kOBQKoKCjZEYjeiyr9fXKf9u8vrXf+tsSP0zkNPSpLKXntb7TomKqF7VwVGDdGmZ9bqwJ5yHdj7iTY9s1aB0efWux2EV25uviZNGiMz0+DBZ2nv3gp9+OGu4+YNHnyWTj65y3HjiYkJx15/9tl+mVlI640FDfXSvo92a9u69ao8fOS4dYc/2ydJatU6XnGt4yXnQlZnKMRHuoBwKCvbqdTUbseWU1JO0muvvVvvnPj4eCUlJejjj8tVVrZTgwefVeO9ZWU7w1N4DOvQo5vKS7YfW/6kdLs69OhWx/gOdejRra5NIALKyj5Samr3Y8spKd1UVrazzjCqT1bWUv3hD3/XoUNH9Pzzfw5FmQiyVq007Y3H1Tlwil7PylZZwTuRLqlJGjyDMrMHzGynmb3b0Fxf1fWlofY3t7rnNO69+ArqOobO1X1so+xbX31io5eO/1k0tR9mzBivf/0rV3feOUu33LKo4TfgK3NHj+reszP1h5RhSh7YT1379op0SU3SmEt8iyWNDnEdIZWScpJKSnYcWy4t3ank5K71zjly5IjKyz9V585JjXovmq6idLuSqn0TT0zproptO/XJcePdVLEtZs5YFysKeykra6nS0iYqLW2ikpO7qqTaGW5p6Y6v3A8TJozUk0++0ExV4sscLK/Q1hdei7rL5Q0GlHNujaTdYaglZAYM6KPi4hJt3lymQ4cOKydntTIyhtaYk5ExVEuWPC1JWrbsOZ133gCZmTIyhionZ7UOHjykzZvLVFxcooED+0ZiN2LK+089r37BJ/R6DOqvg+UV+nT7R9qY95J6jhyidh0T1a5jonqOHKKNeS9FuNrmEa29NGPGeBUWZquwMFuZmcP10EMr5JzTq6+uV1JSQpMu7xUX//vY63/84yX16nVKKEqGpPZdOqltUgdJUny7tjr9/P/Urvei66nJZrsHZWbTJE2TpFNO6d7A7PCKj4/XwoXXadSoWaqsrNSUKRnq2/cM3XjjPUpPP1MZGcN0+eVjddllNyoQyFTnzonKyblNktS37xkaP/589ekzTvHxccrKmqO4uLgI75H/Ls7+vU4bPlDtu3TSNSX5emHeArVqXfXX7Y17c1S8Il+9xgzTrI3P6PC+/cr96a8lSQf2lGvNzX/SFa8vkySt+W2WDuyp/wZxrPG5jyRpzJjvasWKtQoEMtW+fTs9+OC8Y+vS0iaqsDBbkjRnzt3Kzs7Tvn0HlJIyRlOnjtX8+Vdq4cKlevbZArVuHa9OnTpoyZL5EdqT6NFQL329WxdNW/eY2iYmyB09qsFXT1ZWnzFKOPkkZS65Q63i4mStTBuWrlLxP16I7M40kdV1Tfm4SWanSXraOfftxmw0Pb2PW7furydWGZrkJpsY6RJanHu1VdvcgSbdgGlKL9FH4UcfRcZ8ffCGcy699niLeMwcABB9CCgAgJca85j5w5JekdTbzErN7PLQlwXEHnoJaJoGH5Jwzv0oHIUAsY5eApqGS3wAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAALxFQAAAvEVAAAC8RUAAAL5lzrvk3avaRpK3NvuHQ6yJpV6SLaGGi+Zif6pzrGqqNR3EfSdH9c41W0XzM6+ylkARUtDKzdc659EjX0ZJwzGMTP9fwi8VjziU+AICXCCgAgJcIqJrui3QBLRDHPDbxcw2/mDvm3IMCAHiJMygAgJcIKACAlwioIDMbbWbvm9lGM/tVpOuJdWb2gJntNLN3I10Lmg99FH6x3EsElCQzi5OUJekCSX0k/cjM+kS2qpi3WNLoSBeB5kMfRcxixWgvEVBVBkra6Jzb5Jw7JClH0tgI1xTTnHNrJO2OdB1oVvRRBMRyLxFQVXpIKqm2XBocA9B49BGaFQFVxeoY4/l7oGnoIzQrAqpKqaTUasspkrZFqBYgWtFHaFYEVJXXJfUys9PNrI2kCZKeinBNQLShj9CsCChJzrkjkmZKypP0T0lLnXMbIltVbDOzhyW9Iqm3mZWa2eWRrgknhj6KjFjuJX7VEQDAS5xBAQC8REABALxEQAEAvERAAQC8REABALxEQAEAvERAAQC89P/XT7fP40bX+QAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.0\n",
      "D (based on normalized Frobenius) = 0.16061894595623016\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.9345358385928124\n",
      "D (based on normalized Frobenius) = 0.3255375027656555\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.20216560608923156\n",
      "D (based on normalized Frobenius) = 0.40779227018356323\n"
     ]
    }
   ],
   "source": [
    "prob_Yzt = Ro_model.get_ann_confusionM() # estimate of conf_matrix_I\n",
    "\n",
    "compare_set_conf_ma(prob_Yzt, true_set_conf_ma = conf_matrix_I, text=True, n_samp=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Individual D_JS =  0.09656336411544873\n",
      "Individual D_NormF =  0.2345352100180875\n"
     ]
    }
   ],
   "source": [
    "print(\"Individual D_JS = \",Individual_D(conf_matrix_I, prob_Yzt, D=D_JS))\n",
    "print(\"Individual D_NormF = \",Individual_D(conf_matrix_I, prob_Yzt, D=D_NormF))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Negatives values on confusion matrix?\n",
    "> Because the Rodrigues implementation does not consider a probabilistic confusion matrix\n",
    "\n",
    "The model with pre-init over confusion matrices:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot 5 random matrices from the set\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.015766279759775344\n",
      "D (based on normalized Frobenius) = 0.10996438562870026\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.002616200756732275\n",
      "D (based on normalized Frobenius) = 0.04727362096309662\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.00214831019843348\n",
      "D (based on normalized Frobenius) = 0.04796246811747551\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 0.006215842650375265\n",
      "D (based on normalized Frobenius) = 0.0627715066075325\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "D (based on Jensen Shannon) = 1.0668088871574245e-06\n",
      "D (based on normalized Frobenius) = 0.0011573313968256116\n"
     ]
    }
   ],
   "source": [
    "prob_Yzt = Ro_model.get_ann_confusionM(norm=\"softmax\") \n",
    "\n",
    "compare_set_conf_ma(prob_Yzt, true_set_conf_ma = conf_matrix_I, text=True, n_samp=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Individual D_JS =  0.0417849964771893\n",
      "Individual D_NormF =  0.12607141418814438\n"
     ]
    }
   ],
   "source": [
    "print(\"Individual D_JS = \",Individual_D(conf_matrix_I, prob_Yzt, D=D_JS))\n",
    "print(\"Individual D_NormF = \",Individual_D(conf_matrix_I, prob_Yzt, D=D_NormF))"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [conda env:newpy3_tf1]",
   "language": "python",
   "name": "conda-env-newpy3_tf1-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}