-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathEvaluate.py
180 lines (161 loc) · 6.93 KB
/
Evaluate.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
171
172
173
174
175
176
177
178
179
from DictionaryCreate import dicDrain
from FMM import FMM
from BMM import BMM
from unigram import unigram
from HMMC2 import characterHMM
from CBGM import CBGM
from PostProcess import CBGM_poct
def get_answer(rawfile, outfilename, testfilename):
# 吃一个原始语料文件rawfile，输出分词结果格式到outfilename
with open(rawfile, "r", encoding='gbk') as file, \
open(outfilename, "w", encoding='gbk') as answer, \
open(testfilename, 'w', encoding='gbk') as test:
for line in file:
line = line.strip('\n')
sp = line.split(' ')
for pair in sp:
wordandtype = pair.split('/')
if wordandtype[0] == '':
continue
answer.write(wordandtype[0].strip('[') + '/ ') # 处理专有名词
test.write(wordandtype[0].strip('['))
answer.write('\n')
test.write('\n')
def words_pos_set(line):
# 把一行分词结果中所有词的起止下标做成一个集合
line.strip('\n')
words = line.split('/ ')
begin = 0
pos_set = set()
for word in words:
if len(word) == 0:
continue
pos_set.add((begin, begin + len(word) - 1))
begin += len(word)
# print(pos_set)
return pos_set
def evaluation(resultfile, answerfile, analfile):
TT, rtotal, atotal = 0, 0, 0
with open(resultfile, "r", encoding='gbk') as result, \
open(answerfile, "r", encoding='gbk') as answer, \
open(analfile, 'w', encoding='gbk') as anal:
for r_line, a_line in zip(result, answer):
if r_line == '\n':
continue
rset = words_pos_set(r_line)
aset = words_pos_set(a_line)
rtotal += len(rset)
atotal += len(aset)
TT += len(rset & aset)
if len(rset & aset) / len(rset) < 0.6:
anal.write('=========\n')
anal.write(r_line + '\n')
anal.write(a_line + '\n')
anal.write(str(len(rset)) + ' ' + str(len(aset)) + ' ' + str(len(rset & aset)) + '\n')
precison = TT / rtotal
recall = TT / atotal
F1 = 2 * precison * recall / (precison + recall)
print("precison: " + str(precison) + '\n' + "recall: " + str(recall))
print("F1 score: " + str(F1))
print('\n')
return precison, recall, F1
def participation(k, testfile, answerfile, rawfile):
# k为折数 size为每折大小
# 把无标注语料 & 对应分词结果 & 抽词典用原始语料 文件分成十折
with open(testfile, "r", encoding='gbk') as testf, \
open(answerfile, "r", encoding='gbk') as answerf, \
open(rawfile, "r", encoding='gbk') as rawdicfile:
test_list = []
ans_list = []
raw_list = []
for line in testf:
test_list.append(line)
for line in answerf:
ans_list.append(line)
for line in rawdicfile:
raw_list.append(line)
for i in range(k):
with open('/Users/khador/PycharmProjects/NLP-E1/Parts/test' + str(i) + '.txt', "w", encoding='gbk') as part, \
open('/Users/khador/PycharmProjects/NLP-E1/Parts/ans' + str(i) + '.txt', "w",
encoding='gbk') as ans, \
open('/Users/khador/PycharmProjects/NLP-E1/Parts/rawdic' + str(i) + '.txt', "w",
encoding='gbk') as raw:
j = i
while j < len(test_list):
part.write(test_list[j])
ans.write(ans_list[j])
raw.write(raw_list[j])
j += 10
"""for j in range(size):
# testfile.readline()
part.write(testf.readline())
ans.write(answerf.readline())
raw.write(rawdicfile.readline())
j += 1"""
i += 1
def concator(testlist):
# 输入一个用于训练的下标list
with open('./Parts/concatrd.txt', 'w', encoding='gbk') as crd:
# open('concattest.txt', 'w', encoding='gbk') as:
for index in testlist:
with open('./Parts/rawdic' + str(index) + '.txt', 'r', encoding='gbk') as file:
# open('test'+str(index)+'.txt', 'r', encoding='gbk'):
crd.write(file.read())
def ten_fold(method):
resultfilename = './PartDics/' + method + '_10foldresult.txt'
tpresicion, trecall, tF1 = 0, 0, 0
for i in range(0, 10):
testfilename = './Parts/test' + str(i) + '.txt'
print('Fold ' + str(i + 1))
list = [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]
list.remove(i)
concator(list)
dicDrain('./Parts/concatrd.txt', './PartDics/folddic.txt')
if method == 'fmm':
FMM('./PartDics/folddic.txt', testfilename, resultfilename)
if method == 'bmm':
BMM('./PartDics/folddic.txt', testfilename, resultfilename)
if method == 'unigram':
unigram('./PartDics/folddic.txt', testfilename, resultfilename)
if method == 'hmm':
characterHMM('./Parts/concatrd.txt', testfilename, resultfilename)
if method == 'cbgm':
CBGM('./Parts/concatrd.txt', testfilename, resultfilename)
if method == 'cbgmpoct':
CBGM_poct('./Parts/concatrd.txt', testfilename, resultfilename)
p, r, f = evaluation(resultfilename, './Parts/ans' + str(i) + '.txt', 'analtemp.txt')
tpresicion += p
trecall += r
tF1 += f
# print('fold '+str(i)+' is done!')
tpresicion /= 10
trecall /= 10
tF1 /= 10
print('10 fold result:')
print("precison: " + str(tpresicion) + '\n' + "recall: " + str(trecall))
print("F1 score: " + str(tF1))
def acc_FBMM():
# 分析FMM，BMM性能使用
list = []
for i in range(10):
print('acc' + str(i + 1))
list.append(i)
concator(list)
dicDrain('./Parts/concatrd.txt', './PartDics/folddic.txt')
FMM('./PartDics/folddic.txt', './Parts/01test9.txt', 'Facctemp.txt')
BMM('./PartDics/folddic.txt', './Parts/01test9.txt', 'Bacctemp.txt')
print('FMM:')
evaluation('Facctemp.txt', './Parts/01ans9.txt', 'fanaltemp.txt')
print('BMM:')
evaluation('Bacctemp.txt', './Parts/01ans9.txt', 'Banaltemp.txt')
if __name__ == '__main__':
# get_answer('199802.txt',"02_answer.txt",'02_test.txt')
# a = words_pos_set("19980101-01-001-002/ 中共中央/ 总书记/ 、/ 国家/ 主席/ 江/ 泽民/ ")
# b = words_pos_set("19980101-01-001-002中共中央总书/ 记/ 、/ 国家/ 主席/ 江泽/ 民/ ")
# print(a&b)
evaluation('CBGMtestrun.txt', 'gbk_seg_answer.txt','CBGManalysis.txt')
# evaluation('gbk_seg_answer.txt','lhx199801_seg&pos_true.txt')
# participation(10, '199801_sent.txt',"gbk_seg_answer.txt",'199801_seg&pos.txt')
# concator([0, 1, 2, 3, 4, 5, 6, 7, 8])
# ten_fold('hmm')
# acc_FBMM()