-
Notifications
You must be signed in to change notification settings - Fork 0
/
Learn.py
170 lines (164 loc) · 6.04 KB
/
Learn.py
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
# -*- coding: gb2312 -*-
# Learn from the training data
# Classify partited data
# sentences description type(desc_type):action(0), speech(1), description(2)
# Character Personality type(Personality):
# [impulsive(1)/calm(0),Extrovert(1)/Introvert(0),
# Optimistic(1)/pessimistic(0)]
import json
import math
import glob
# Data Structures
descTypeProb = dict() # {word:[probability of 3 desc_type]}
# {word:{desctype:6 probabilities for each personalites},...}
personalityProb = dict()
sentenceTypePrior = [0,0,0]#[0]:quote [1]: normal [2]:half
personalityPrior = [0,0,0,0,0,0]
wordCount = 0
characterCount = 0
tagFiles = 'train/classifier training data/'
partitionFiles = 'train/classifier training data/*.json'
modelFilePath = 'model.txt'
count0 = 0
count1=0
count2 = 0
# parsing tag data
def parseTagFile(path):
tagFile = open(tagPath,'r')
tagDict = dict()
while True:
line = tagFile.readline()
if line:
charName = line.strip('\n').decode('utf-8')
#print charName
line = tagFile.readline()
tags = line.strip('\n').split()
if tags[0] == '\xe5\x86\xb2\xe5\x8a\xa8': # Impulsive
tags[0] = 1;
else:
tags[0] = 0
if tags[1] == '\xe5\xa4\x96\xe5\x90\x91': # Extrovert
tags[1] = 1;
else:
tags[1] = 0
if tags[2] == '\xe4\xb9\x90\xe8\xa7\x82': # Optimistic
tags[2] = 1;
else:
tags[2] = 0
#print tags
tagDict.setdefault(charName,[])
tagDict[charName] = tags
#print tagDict
else:
break
tagFile.close()
return tagDict
# load training data
partitionPath = glob.glob(partitionFiles)
for p in partitionPath:
learnFile = open(p, 'r')
novelName = p[15:len(p)-9]
tagPath = tagFiles + novelName + '.tag'
tagDict = parseTagFile(tagPath)
learnFileText = learnFile.read()
dataSets = json.loads(learnFileText.decode('utf-8'))
learnFile.close()
# training
for dataSet in dataSets:
cname = ''
for name in tagDict.keys():
if name not in dataSet['chars']:
continue
else:
cname = name
print name
characterCount +=1
for i in range(0,3):
if tagDict[cname][i]==1:
personalityPrior[2*i]+=1
#personality = dataSet['Personality']
for i in dataSet['sentences']:
descType = i['type']
if i['type'] == 0:
count0 += 1
if descType == 0:
descType = 0 #quote
sentenceTypePrior[0] += 1
elif descType == 3:
descType = 1 #normal
sentenceTypePrior[1] += 1
else:
sentenceTypePrior[2] += 1#half
descType = 2
if descType == 1:
count1 +=1
if descType ==2:
count2 +=1
for j in i['tokens']:
word = j['word']
wordCount += 1
descTypeProb.setdefault(word, [0, 0,0])
descTypeProb[word][descType] += 1
if cname != '':
personalityProb.setdefault(word, {})
for k in range(0, 3):
personalityProb[word].setdefault(k, [0, 0, 0, 0, 0, 0])
for i in range(0, 3):
if tagDict[cname][i] == 1:
personalityProb[word][descType][2 * i] += 1
else:
personalityProb[word][descType][2 * i + 1] += 1
# smoothing & modeling
#priors
print count0,count1,count2
c0 = sentenceTypePrior[0]
c1 = sentenceTypePrior[1]
c2 = sentenceTypePrior[2]
characterCount += 8 #smoothing
sentenceTypePrior[0] = math.log((c0+1)/float(c0+c1+c2+3))
sentenceTypePrior[1] = math.log((c1+1)/float(c0+c1+c2+3))
sentenceTypePrior[2] = math.log((c2+1)/float(c0+c1+c2+3))
personalityPrior[1] = math.log((characterCount-personalityPrior[0]-4)/\
float(characterCount))
personalityPrior[0] = math.log((personalityPrior[0]+4)/\
float(characterCount))
personalityPrior[3] = math.log((characterCount-personalityPrior[2]-4)/\
float(characterCount))
personalityPrior[2] = math.log((personalityPrior[2]+4)/\
float(characterCount))
personalityPrior[5] = math.log((characterCount-personalityPrior[4]-4)/\
float(characterCount))
personalityPrior[4] = math.log((personalityPrior[4]+4)/\
float(characterCount))
modelFile = open(modelFilePath, 'w')
for i in range (0,3):
modelFile.write(str(sentenceTypePrior[i])+ '\t')
modelFile.write('\n')
for i in range (0,6):
modelFile.write(str(personalityPrior[i])+ '\t')
modelFile.write('\n')
modelFile.write(str(len(descTypeProb.keys())) + '\n')
for word in descTypeProb.keys():
modelFile.write(word.encode('utf-8') + '\t')
for k in range(0, 3):
descTypeProb[word][k] = (descTypeProb[word][k] + 1) / \
float(3 * len(descTypeProb.keys()) + wordCount)
descTypeProb[word][0] *=2
descTypeProb[word][1] /=2
#descTypeProb[word][2] *=2
modelFile.write(str(math.log(descTypeProb[word][k])) + '\t')
# print word, descTypeProb[word]
modelFile.write('\n')
for word in personalityProb.keys():
modelFile.write(word.encode('utf-8') + '\t')
for k in range(0, 3):
for p in range(0, 6):
personalityProb[word][k][p] = (personalityProb[word][k][p] + 1) / \
float(18 * len(personalityProb.keys()) + 3 * wordCount)
modelFile.write(str(math.log(personalityProb[word][k][p])) + '\t')
# print word, personalityProb[word]
modelFile.write('\n')
modelFile.close()
# print dataSet.keys()
# print type(dataSet['sentences'])
# print dataSet['sentences'][1]['tokens'][1]['word']