def main():
if len(sys.argv) != 4:
print 'Invalid Command. Should be python MainHandler train_file classifier_file type(1:Neutral,2:Polarity) '
sys.exit(0)
print sys.argv[1]
pdata = bayesianClassifier.NBayesianMethod(sys.argv[1], sys.argv[3])
pdata.trainClassifier(sys.argv[2])
def main():
if len(sys.argv) != 2:
print 'Invalid Command. Should be python MainHandler test_file '
sys.exit(0)
f = open('neutral.pickle')
classifier = pickle.load(f)
f.close()
pdata = bayesianClassifier.NBayesianMethod(sys.argv[1], "1")
alltweets = []
print 'Accuracy:', nltk.classify.accuracy(classifier,
[(pdata.features(d), s)
for (d, s) in pdata.tweets])
for (t, s) in pdata.tweets:
p = classifier.classify(pdata.features(t))
alltweets.append((t, p, s))
f = open('polar.pickle')
classifier = pickle.load(f)
f.close()
pdata = bayesianClassifier.NBayesianMethod(sys.argv[1], 2)
print 'Accuracy:', nltk.classify.accuracy(classifier,
[(pdata.features(d), s)
for (d, s) in pdata.tweets])
print 'Tweet \t\t\t Actual \t\t Predicted'
wrong = 0
for (t, s) in pdata.tweets:
p = classifier.classify(pdata.features(t))
alltweets.append((t, p, s))
for (t, p, s) in alltweets:
if p != s:
print p, s
print t, '\t', s, '\t', p
wrong = wrong + 1
print wrong
print "Accuracy:", float(wrong) / len(pdata.tweets)
def main():
if len(sys.argv) != 4:
print 'Invalid Command. Should be python MainHandler test_file classifier_file type(1:neutral,2:polar)'
sys.exit(0)
f = open(sys.argv[2])
classifier = pickle.load(f)
f.close()
pdata = bayesianClassifier.NBayesianMethod(sys.argv[1], sys.argv[3])
print 'Accuracy:', nltk.classify.accuracy(classifier,
[(pdata.features(d), s)
for (d, s) in pdata.tweets])
print 'Tweet \t\t\t Actual \t\t Predicted'
import sys
import bayesianClassifier
import pickle
import nltk
f = open('neutral.pickle')
nc = pickle.load(f)
f.close()
ffeatred = open('neutral_woswrds.pickle')
ncred = pickle.load(ffeatred)
ffeatred.close()
f = open('polar.pickle')
pc = pickle.load(f)
f.close()
pdata = bayesianClassifier.NBayesianMethod('ltest.csv')
if (i <= 100):
tweets.append(tweet['text'])
#print( '@%s tweeted: %s' % ( tweet['user']['screen_name'], tweet['text'] ) )
i = i + 1
else:
break
print 'Total Tweets:', len(tweets)
f = open('neutral.pickle')
nc = pickle.load(f)
f.close()
posTweets = []
negTweets = []
nutrl = []
ppolar = bayesianClassifier.NBayesianMethod()
print "************* Neutral Tweets ********************"
for t in tweets:
#print t
s = nc.classify(ppolar.features(t))
#print t,s
if (int(s) != 2):
ppolar.tweets.append(t)
else:
nutrl.append(t)
print t
f = open('polar.pickle')
pc = pickle.load(f)
f.close()
print "**************************"