-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtest
199 lines (183 loc) · 7.07 KB
/
test
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
# -*- coding: utf-8 -*-
"""
Created on Mon Oct 22 16:10:52 2018
@author: Administrator
"""
import math
class Featureselect():
def __init__(self,inputx,inputy):
self.inputx=inputx #训练数据集
self.inputy=inputy
self.feature=[] # x的特征
self.feaNum_dict={} #存储的是x的特征以及对应的训练集的下标
self.feaX_dict={}
self.D_entro=0 #存的是训练集的信息熵
self.Dik_list=[] #存储的是H(D|A)
self.entropyDice={}
self.y_class_dict={} #存储y的可能取值，即Ck
self.g=[]#存储的是各个特征的信息增益
self.gr=[]#存储的是信息增益比
##################3以下是对函数的调用
self.select_y()
self.select_x()
self.D_entropy()
self.Dik_entropy()
# self.count_entropy()
def select_y(self):# 生成y的可能取值的集合，Ck
y_class=[]
for y in self.inputy:
if y not in y_class:
y_class.append(y)
else:
pass
flag=0
y_num=[]
for yi in y_class:
for y in self.inputy:
if yi==y:
flag+=1
else:
pass
y_num.append(flag)
flag=0
self.y_class_dict=dict(zip(y_class,y_num)) #y的类，和每个类的训练集个数
# print('self.y_class_dict',self.y_class_dict)
def select_x(self):
feature=[]
feature1=[]
for i in range(len(self.inputx[0])):
for x in self.inputx:
if x[i] not in feature:
feature.append(x[i])
else:
pass
feature1.append(feature) #双重list
feature=[]
# print('feature1',feature1)
flag=0
xi_num=[]
Xi_x=[]
num=[]
x_x=[]
flag
for index,xi_feature_list in enumerate(feature1): #取出xi的所有特征的集合
for feature in xi_feature_list: #重特征集合中取出一个特征
index_list=[]
x_list=[]
for index1,x in enumerate(self.inputx):
if x[index]==feature:
index_list.append(index1) #将属于这个特征的x的下标保存下来
x_list.append(x)
flag+=1
else:
pass
xi_num.append(index_list)
Xi_x.append(x_list)
# flag=0
num.append(xi_num)
x_x.append(Xi_x)
xi_num=[]
Xi_x=[]
# print('num',num)
dict_list=[]
xx_list=[]
# print('x',x_x)
for index,feature in enumerate(feature1):
f_num_dict = dict(zip(feature,num[index]))
f_x_dict=dict(zip(feature,x_x[index]))
# print('f_num_dict',f_num_dict)
dict_list.append(f_num_dict)
xx_list.append(f_x_dict)
for i in range(len(self.inputx[0])):
self.feaNum_dict[i]=dict_list[i]
self.feaX_dict[i]=xx_list[i]
# print('self.feaNum_dict',self.feaNum_dict)
# print('self.feaX_dict',self.feaX_dict)
# for index,x_feature in enumerate(x):
def D_entropy(self):#计算训练集经验熵
y_pro=[]
for Ck in self.y_class_dict.keys():
y_pro.append(self.y_class_dict[Ck]/len(self.inputy))
y_entropy=0
for pro in y_pro:
# print('pro',pro)
y_entropy=y_entropy - pro * math.log(pro,2)
self.D_entro=y_entropy
# print('self.D_entro',self.D_entro)
# =====#========================================================================
# def count_entropy(self):#计算信息熵即H(D|A)
# entropyDice={}
# for Xi in self.feaNum_dict.keys(): #指第i个特征
# num_sum=0
# for value in self.feaNum_dict[Xi].values(): #指第i个特征可能的取值
# num_sum=num_sum-value*math.log(value,2) #以2为低
# entropyDice[Xi]=num_sum
# self.entropyDice=entropyDice
# print('entropyDice',entropyDice)
# =============================================================================
def Dik_entropy(self):#计算经验条件熵
Dik_entro=[]
num_long=len(self.inputx) #输入的个数
for Xi in self.feaNum_dict.keys():#Xi是输入的第几个x
Dik=[]
for Di in self.feaNum_dict[Xi].keys():
# print('self.feaNum_dict[Xi][Di]',self.feaNum_dict[Xi][Di])
long=len(self.feaNum_dict[Xi][Di])
pro=long/num_long # Di/D 的概率
# print('pro',pro,long,num_long)
Dik_ent=0
for y in self.y_class_dict.keys(): #y的类别
flag=0
for index in self.feaNum_dict[Xi][Di]:
if self.inputy[index]==y:
flag+=1
else:
pass
Di_pro=flag/long
# print('Di_pro',Di_pro,flag,long)
try:#有可能Di_pro为0
Dik_ent=Dik_ent-Di_pro*math.log(Di_pro,2)
except:
pass
# print('Dik_ent',Dik_ent,Dik_ent*pro)
Dik.append(Dik_ent*pro)
Dik_entro.append(Dik)
# print(Dik_entro)
for D_list in Dik_entro:
num_sum=0
for num in D_list:
num_sum=num_sum+num
self.Dik_list.append(num_sum)
# print('self.Dik_list',self.Dik_list)
def G_infor(self):#计算信息增益
for Hi in self.Dik_list:
self.g.append(self.D_entro-Hi)
# print('self.g',self.g)
return self.g
def GR_infor(self):#计算信息增益比
num_long=len(self.inputx)
for X in self.feaNum_dict.keys():
num_sum=0
for value in self.feaNum_dict[X].values():
long=len(value)#就是X特征这个取值的个数了
num_sum=num_sum-(long/num_long)*math.log(long/num_long,2)
self.gr.append(num_sum)
# print('self.gr',self.gr)
return self.gr
if __name__=='__main__':
x1={1:'青年',2:'中年',3:'老年'} #年龄
x2={1:'否',2:'是'}#有工作
x3={1:'否',2:'是'}#有自己的房子
x4={1:'一般',2:'好',3:'非常好'}#信贷情况
inputx=[[1, 1, 1, 1], [1, 1, 1, 1], [2, 1, 1, 1], [3, 1, 1, 1]]
inputy=[1, 1, 1, 1]
# inputx=[[1,1,1,1],[1,1,1,2],[1,2,1,2],[1,2,2,1],[1,1,1,1],
# [2,1,1,1],[2,1,1,2],[2,2,2,2],[2,1,2,3],[2,1,2,3],
# [3,1,2,3],[3,1,2,2],[3,2,1,2],[3,2,1,3],[3,1,1,1]]
y={1:'是',2:'否'}
inputy=[1,1,2,2,1,1,1,2,2,2,2,2,2,2,1]
a=Featureselect(inputx,inputy)
a.G_infor()
a.GR_infor()
# a.Dik_entropy()
# a.count_entropy()