def classify(self, source, kaggle = None):
data = parseReviews.getReviewList(source, defaultToZero = False)
predictions = []
for review in data:
tags = []
for entry in review[1]:
probs = dict()
for sentiment in self.sentiment_probabilities:
probs[sentiment] = math.log(self.sentiment_probabilities[sentiment])
for word in entry[1]:
if word in self.word_probabilities:
for p in self.word_probabilities[word]:
probs[p] += math.log(self.word_probabilities[word][p])
tags.append(max(probs.iteritems(), key=operator.itemgetter(1))[0])
print tags
predictions.append(tags)
if kaggle != None:
with open(kaggle,'w') as f:
f.write('Id,answer\n')
i = 0
for seq in predictions:
for tag in seq:
f.write(str(i) + ',' + str(tag) + '\n')
i += 1
return predictions
def get_train_data(filename, destination):
#with open(destination, 'w') as d:
with open(filename) as f:
result = dict()
lst = parseReviews.getReviewList(filename)
for i in lst:
linelst = i[1]
for j in linelst:
result[j[1].strip()] = ToSentiment(j[0])
#d.write(j[1].strip() + " : " + ToSentiment(j[0]) + "\n")
#d.close()
return result
def get_test_data(filename, destination):
with open(destination, 'w') as d:
with open(filename) as f:
result = []
lst = parseReviews.getReviewList(filename)
for i in lst:
linelst = i[1]
for j in linelst:
result.append(j[1].strip())
d.write("Id,answer"+"\n")
i = 0
for sample in result:
d.write(str(i) +"," + str(parseReviews.strToSentiment(classifier.classify(gen_bow(sample)))) + "\n")
i = i +1
d.close()
return result
def get_train_data(filename, destination):
with open(destination, 'w') as d:
with open(filename) as f:
result = dict()
print 'inside script'
lst = parseReviews.getReviewList(filename)
d.write('Id,Answer' + '\n')
print 'writing to file'
for i in lst:
linelst = i[1]
k = 0
for j in linelst:
result[j[1].strip()] = j[0]
d.write(str(k)+ "," + str(j[0]) + "\n")
k = k+1
d.close()
return result
def parse_training(self, source):
data = parseReviews.getReviewList(source, defaultToZero = False)
for review in data:
for entry in review[1]:
self.update_counts(entry)
def parseTraining(filename): with open(filename) as f: a = parseReviews.getReviewList(filename) dic = parseReviews.wordSentimentMapBasic(a) return list(dic.keys())
