Sentiment Analysis with Python- SentiWordNet

Here is a code for sentiment analysis explained with sentiwordnet. The code is here for determining the sentiment analysis of a sentence.

I have used this on the climate dataset with me. And found the most positive sentiment-oriented documents and most negatively oriented documents, also the most lengthy documents too.

Here is a simple code of sentiment analysis using sentiWordNet.

Initialization as follows:

Now, use the sentiwordnet package and find sentiments.

…

This returns the positive and negative scores of the sentence.

The output is as follows:

…

The full code is here:

import nltk
from nltk.corpus import wordnet as wn
from nltk.corpus import sentiwordnet as swn

nltk.download('sentiwordnet')
nltk.download('wordnet')
 
sentence='The painting is beautiful and serene.'

tokenSentence = nltk.word_tokenize(sentence)
print(tokenSentence)
posTagSentence = nltk.pos_tag(tokenSentence)
print(posTagSentence)
 
sentimentPositive = 0.0
sentimentNegatve = 0.0
sentenceCount = 0

for wordTagPair in posTagSentence:

  word = wordTagPair[0]
  posTag = wordTagPair[1]
  print(word, posTag)

  if posTag.startswith('J'):
    posTag =  wn.ADJ
  elif posTag.startswith('R'):
    posTag =  wn.ADV    
  elif posTag.startswith('N'):
    posTag =  wn.NOUN
  else:
      continue
  
  wordSynst = wn.synsets(word, pos=posTag)
  if not wordSynst:
    continue  
  chosenSynst = wordSynst[0]
  sentiWordNet = swn.senti_synset(chosenSynst.name())
  print(sentiWordNet)
  sentimentPositive += sentiWordNet.pos_score() 
  sentimentNegatve +=  sentiWordNet.neg_score()
  sentenceCount += 1

print('The positive and negative sentiments of given sentence are: ')
print (sentimentPositive, sentimentNegatve)