-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathRuleMakerNextWord
158 lines (124 loc) · 3.98 KB
/
RuleMakerNextWord
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
import nltk
from nltk import *
from nltk.corpus import brown
import numpy
from numpy import *
import time
def findmaxtag(good, bad):
curMax = 0
curTag = ' '
for tag in good:
if(good[tag]-bad[tag] > curMax):
curMax = good[tag]-bad[tag]
curTag = tag
return curTag, curMax
def applyrule_previoustag(from_tag, to_tag, previous_tag, oldcorpus):
i = 0
newcorpus = []
for word in oldcorpus:
if (i==0):
newcorpus.append(tuple([word[0], word[1]]))
elif (oldcorpus[i-1][1] == previous_tag and word[1] == from_tag):
newcorpus.append(tuple([word[0], to_tag]))
print 'Orig: ',word, 'New: ',newcorpus[-1]
else:
newcorpus.append(tuple([word[0],word[1]]))
return newcorpus
def comparetags(true,test):
if(len(true) != len(test)):
return 'Cant compare....lengths not the same'
i,wrong = 0,0
for i in range(len(true)):
if(test[i][1] != true[i][1]):
wrong += 1
return (len(true)-wrong)*100/len(true)
#####################################
start = time.time()
trainpercent = 50
words = brown.words()
ntotal = len(words)
words = words[:int32(ntotal*trainpercent/100)]
testwords = words[-(ntotal-len(words)):]
t = brown.tagged_words(simplify_tags=True)[:len(words)]
words = [w.lower() for w in words]
tagwords = []
for word in t:
tagwords.append(tuple([word[0].lower(),word[1]]))
nwords = len(tagwords)
tagSet = []
for word in tagwords:
tagSet.append(word[1])
tagSet = set(tagSet)
tagSet = list(tagSet)
fdist = FreqDist(words)
fkeys = fdist.keys()
fprob = []
for i in range(0,fdist.B()):
fprob.append(fdist.freq(fkeys[i]))
tagcfd = ConditionalFreqDist(tagwords)
tagcfdkeys = tagcfd.keys()
mostlikelytag = []
for i in range(nwords):
w = tagwords[i][0]
mostlikelytag.append(tuple([w, tagcfd[w].max()]))
##wrong = 0
##for i in range(nwords):
## if (mostlikelytag[i][1] != tagwords[i][1]):
## wrong = wrong + 1
accuracy = comparetags(tagwords, mostlikelytag)
elapsed = time.time()-start
print 'Time: ',elapsed,' Accuracy: ',accuracy
#####################################
ambclass = dict()
for word in words:
if (word not in ambclass):
tag = tagcfd[word].items()
tmp = []
for i in tag:
tmp.append(i[0])
ambclass[word] = tmp
#####################################
num_good_T = dict()
num_bad_T = dict()
for itr in tagSet:
num_good_T[itr] = 0
num_bad_T[itr] = 0
bestrulelist = []
thresh = 100
startOverall = time.time()
while (True):
start = time.time()
maxval = 0
counter1 = 0
for from_tag in tagSet:
print counter1
counter1 += 1
counter2 = 0
for to_tag in tagSet:
print " ", counter2
counter2 += 1
i = 1
for word in mostlikelytag[1:]:
correct_tag = tagwords[i][1]
taglist = ambclass[word[0]]
if(to_tag in taglist):
if(from_tag != to_tag):
if(correct_tag == to_tag and word[1] == from_tag):
num_good_T[mostlikelytag[i-1][1]] += 1
elif(correct_tag == from_tag and word[1] == from_tag):
num_bad_T[mostlikelytag[i-1][1]] += 1
i += 1
maxtag, val = findmaxtag(num_good_T,num_bad_T)
if (val>maxval):
maxval = val
changefrom_tag = from_tag
changeto_tag = to_tag
changeprev_tag = maxtag
if (maxval < thresh):
break
mostlikelytag = applyrule_previoustag(changefrom_tag,changeto_tag,changeprev_tag,mostlikelytag)
bestrulelist.append([changefrom_tag,changeto_tag,changeprev_tag])
accuracyRule_1 = comparetags(tagwords, mostlikelytag)
elapsed = time.time()-start
print 'Time: ',elapsed,' Accuracy of rule: ',accuracy
print 'TOTAL TIME: ', time.time()-startOverall