def polarity(tweet,sentDict): # polarity aggregate of a tweet from sentiWordnet dict
pos=0.0
neg=0.0
neu=0.0
n_words=0
for w in tweet.split():
if w in sentDict.keys():
n_words=n_words+1
pos=pos+sentDict[w][1]
neg=neg+sentDict[w][2]
neu=neu+sentDict[w][3]
if features.hashTest(w) and w[1:] in sentDict.keys():
pos=pos+2*sentDict[w[1:]][1] # more weight for hashed words
neg=neg+2*sentDict[w[1:]][2]
neu=neu+2*sentDict[w[1:]][3]
if (n_words ==0 ):
return [pos,neg,neu]
else:
return [pos/n_words,neg/n_words,neu/n_words]
def polarity(tweet,sentDict): # polarity aggregate of a tweet from sentiWordnet dict
pos=0.0
neg=0.0
neu=0.0
n_words=0
for w in tweet.split():
if w in sentDict.keys():
n_words=n_words+1
pos=pos+sentDict[w][1]
neg=neg+sentDict[w][2]
neu=neu+sentDict[w][3]
if features.hashTest(w) and w[1:] in sentDict.keys():
pos=pos+2*sentDict[w[1:]][1] #imp. - more weight for hashed words
neg=neg+2*sentDict[w[1:]][2]
neu=neu+2*sentDict[w[1:]][3]
if (n_words ==0 ):
return [pos,neg,neu]
else:
return [pos/n_words,neg/n_words,neu/n_words]
def posPolarity(tweet,sentDict): # polarity aggregate of a tweet from sentiWordnet dict
pos=0.0
neg=0.0
neu=0.0
posScores={}
posNumber={}
posNumber['a']=0
posNumber['n']=0
posNumber['v']=0
posNumber['r']=0
posScores['a']=0.0
posScores['n']=0.0
posScores['v']=0.0
posScores['r']=0.0
n_words=0
nlpos={}
nlpos['a']=['JJ','JJR','JJS'] # adjective tags in nltk
nlpos['n']=['NN','NNS','NNP','NNPS'] # nouns ...
nlpos['v']=['VB','VBD','VBG','VBN','VBP','VBZ','IN'] # verbs
nlpos['r']=['RB','RBR','RBS'] # adverbs
text=tweet.split()
tags=nltk.pos_tag(text)
for z in tags:
w=("testtesttestt",'mdr')
y=list(z)
if (y[1] in nlpos['a']):
w=(y[0],'a')
posNumber['a']+=1
if (y[1] in nlpos['n']):
w=(y[0],'n')
posNumber['n']+=1
if (y[1] in nlpos['v']):
w=(y[0],'v')
posNumber['v']+=1
if (y[1] in nlpos['r']):
w=(y[0],'r')
posNumber['r']+=1
if w in sentDict.keys():
n_words=n_words+1
posScores[w[1]]+=sentDict[w][0]-sentDict[w][1]
pos=pos+sentDict[w][0]
neg=neg+sentDict[w][1]
neu=neu+sentDict[w][2]
if features.hashTest(list(w)[0]) and (list(w)[0][1:],list(w)[1]) in sentDict.keys():
n_words=n_words+1
posNumber[w[1]]+=1
posScores[w[1]]+=sentDict[(list(w)[0][1:],list(w)[1])][0]-sentDict[(list(w)[0][1:],list(w)[1])][1]
pos=pos+2*sentDict[(list(w)[0][1:],list(w)[1])][0] # more weight for hashed words
neg=neg+2*sentDict[(list(w)[0][1:],list(w)[1])][1]
neu=neu+2*sentDict[(list(w)[0][1:],list(w)[1])][2]
if (n_words ==0 ):
return [pos,neg,neu,posScores['a'],posScores['n'],posScores['v'],posScores['r'],posNumber['a'],posNumber['n'],posNumber['v'],posNumber['r']]
else:
return [pos/n_words,neg/n_words,neu/n_words,posScores['a']/n_words,posScores['n']/n_words,posScores['v']/n_words,posScores['r']/n_words,posNumber['a']/n_words,posNumber['n']/n_words,posNumber['v']/n_words,posNumber['r']/n_words] # do not use the neutral scoore it will f**k up evertg
def posPolarity(tweet,sentDict): # polarity aggregate of a tweet from sentiWordnet dict
pos=0.0
neg=0.0
neu=0.0
posScores={}
posNumber={}
posNumber['a']=0
posNumber['n']=0
posNumber['v']=0
posNumber['r']=0
posScores['a']=0.0
posScores['n']=0.0
posScores['v']=0.0
posScores['r']=0.0
n_words=0
nlpos={}
nlpos['a']=['JJ','JJR','JJS'] # adjective tags in nltk
nlpos['n']=['NN','NNS','NNP','NNPS'] # nouns ...
nlpos['v']=['VB','VBD','VBG','VBN','VBP','VBZ','IN'] # verbs
nlpos['r']=['RB','RBR','RBS'] # adverbs
text=tweet.split()
tags=nltk.pos_tag(text)
for z in tags:
w=("testtesttestt",'mdr')
y=list(z)
if (y[1] in nlpos['a']):
w=(y[0],'a')
posNumber['a']+=1
if (y[1] in nlpos['n']):
w=(y[0],'n')
posNumber['n']+=1
if (y[1] in nlpos['v']):
w=(y[0],'v')
posNumber['v']+=1
if (y[1] in nlpos['r']):
w=(y[0],'r')
posNumber['r']+=1
if w in sentDict.keys():
n_words=n_words+1
posScores[w[1]]+=sentDict[w][0]-sentDict[w][1]
pos=pos+sentDict[w][0]
neg=neg+sentDict[w][1]
neu=neu+sentDict[w][2]
if features.hashTest(list(w)[0]) and (list(w)[0][1:],list(w)[1]) in sentDict.keys():
n_words=n_words+1
posNumber[w[1]]+=1
posScores[w[1]]+=sentDict[(list(w)[0][1:],list(w)[1])][0]-sentDict[(list(w)[0][1:],list(w)[1])][1]
pos=pos+2*sentDict[(list(w)[0][1:],list(w)[1])][0] # imp. - more weight for hashed words
neg=neg+2*sentDict[(list(w)[0][1:],list(w)[1])][1]
neu=neu+2*sentDict[(list(w)[0][1:],list(w)[1])][2]
if (n_words ==0 ):
return [pos,neg,neu,posScores['a'],posScores['n'],posScores['v'],posScores['r'],posNumber['a'],posNumber['n'],posNumber['v'],posNumber['r']]
else:
return [pos/n_words,neg/n_words,neu/n_words,posScores['a']/n_words,posScores['n']/n_words,posScores['v']/n_words,posScores['r']/n_words,posNumber['a']/n_words,posNumber['n']/n_words,posNumber['v']/n_words,posNumber['r']/n_words] # do not use the neutral scoore it will f**k up evertg