def subjectivity_lexicon_features(phrase):
features = {}
# FIXME - MUST disambiguate POS
# Feature Subjectivity Classification
Subj_sentiments = defaultdict(lambda: 0)
for word in phrase:
negated = ('_neg' in word)
if negated: word = word[:-4]
entry = lexSubj.lookup(word)
if entry.prior != '':
label = (entry.type, entry.prior) # ex. ('strongsub','positive')
if negated:
if label[1] == 'positive':
label = ('weaksubj', 'negative')
elif label[1] == 'negative':
label = ('weaksubj', 'positive')
Subj_sentiments[label] += 1
for sentiment, count in Subj_sentiments.items():
features[('Subj-%s-%s_count' % sentiment)] = count
return features
def subjectivity_lexicon_features(phrase):
features = {}
# FIXME - MUST disambiguate POS
# Feature Subjectivity Classification
Subj_sentiments = defaultdict(lambda:0)
for word in phrase:
negated = ('_neg' in word)
if negated: word = word[:-4]
entry = lexSubj.lookup(word)
if entry.prior != '':
label = (entry.type, entry.prior) # ex. ('strongsub','positive')
if negated:
if label[1] == 'positive':
label = ('weaksubj','negative')
elif label[1] == 'negative':
label = ('weaksubj','positive')
Subj_sentiments[label] += 1
for sentiment,count in Subj_sentiments.items():
features[ ('Subj-%s-%s_count' % sentiment) ] = count
return features
def subjectivity_lexicon_features(phrase):
features = {}
# FIXME - MUST disambiguate POS
# Feature Subjectivity Classification
# print '\n\n'
# print phrase
Subj_sentiments = defaultdict(lambda: 0)
for word in phrase:
entry = lexSubj.lookup(word)
# print '\t', word, '\t\t\t', entry
if entry.prior != "":
label = (entry.type, entry.prior) # ex. ('strongsub','positive')
Subj_sentiments[label] += 1
for sentiment, count in Subj_sentiments.items():
# print 'feat: ', 'Subj-%s-%s_count' % sentiment, '\t\t', count
features["Subj-%s-%s_count" % sentiment] = count
return features
def subjectivity_lexicon_features(phrase):
features = {}
# FIXME - MUST disambiguate POS
# Feature Subjectivity Classification
#print '\n\n'
#print phrase
Subj_sentiments = defaultdict(lambda: 0)
for word in phrase:
entry = lexSubj.lookup(word)
#print '\t', word, '\t\t\t', entry
if entry.prior != '':
label = (entry.type, entry.prior) # ex. ('strongsub','positive')
Subj_sentiments[label] += 1
for sentiment, count in Subj_sentiments.items():
#print 'feat: ', 'Subj-%s-%s_count' % sentiment, '\t\t', count
features['Subj-%s-%s_count' % sentiment] = count
return features