#add your keys below
public_key = "PUBaxIIThb7R7A%ZePGtuLDl-olsIsXZ"
private_key = "PRIBfWPexYV6YyqR3AmwIK%QpNIil7-0"
#build the Free API object with your keys
chatterboxapi = SentimentAnalysisFree(public_key, private_key)
f = open('../test set.txt', 'r')
total = 0
count = 0
for line in f:
total += 1
tweet = line.split(' .:. ')
sentiment = tweet[0]
result = chatterboxapi.classifytext("en", tweet)
sentiment_label = result['sent']
if sentiment_label == 1:
sentiment_label = 'pos'
elif sentiment_label == -1:
sentiment_label = 'neg'
else:
sentiment_label = 'neutral'
# score = result['value']
# if abs(score) < 0.20:
# sentiment_label = 'neutral'
# elif score > 0:
# sentiment_label = 'pos'
"Well, Saudi Airlines have really messed up our holiday. Pretty pissed off they can get away with it.",
"How did someone make it to our site using the keywords: \"different types of birth control\"? #odd",
"Apple Mail, why do you crash so? #grrrr",
"If you ever wonder 'how bad can a coffee from burger king be, really?' the answer is bad, very bad. #badcoffeebadmorning",
"Great meeting with TK from the Queen Mary student paper - really interesting questions!"]
highestnumber = 0
highesttext = False
try:
for sampletext in sampletexts:
print "Text: " + sampletext
#Here we do the actual classification. We pass in a language
#identifier and the text we wish to be classified.
classification = chatterboxapi.classifytext("en",sampletext)
#Uncomment this line if you want to inspect the result.
#print classification.body
#Value is the predicted strength of the sentiment in the text
sentiment_value = classification.body['value']
if sentiment_value > highestnumber:
highestnumber = sentiment_value
highesttext = sampletext
#Sent is the sentiment class. 1 is positive, -1 is negative
sentiment_label = classification.body['sent']
if abs(sentiment_value) < 0.25:
"How did someone make it to our site using the keywords: \"different types of birth control\"? #odd",
"Apple Mail, why do you crash so? #grrrr",
"If you ever wonder 'how bad can a coffee from burger king be, really?' the answer is bad, very bad. #badcoffeebadmorning",
"Great meeting with TK from the Queen Mary student paper - really interesting questions!"
]
highestnumber = 0
highesttext = False
try:
for sampletext in sampletexts:
print "Text: " + sampletext
#Here we do the actual classification. We pass in a language
#identifier and the text we wish to be classified.
classification = chatterboxapi.classifytext("en", sampletext)
#Uncomment this line if you want to inspect the result.
#print classification.body
#Value is the predicted strength of the sentiment in the text
sentiment_value = classification.body['value']
if sentiment_value > highestnumber:
highestnumber = sentiment_value
highesttext = sampletext
#Sent is the sentiment class. 1 is positive, -1 is negative
sentiment_label = classification.body['sent']
if abs(sentiment_value) < 0.25:
