A novel intuitionistic fuzzy inference system for sentiment analysis

Sentiment analysis is important area of research in Psychology that helps to predict the attitude or personality trait of human. In the present study, we propose novel intuitionistic fuzzy inference system (IFIS) for the sentiment analysis. The research paper does the sentiment analysis of using tweets and predicts the personality trait characteristics of the tweeting individual through proposed IFIS. Twitter data was analyzed using Natural Language Processing Toolkit (NLTK) through TextBlob in Google Colaboratory for their subjectivity and polarity to predict the score of their positivity using proposed novel IFIS. Performance of proposed IFIS is compared with Mamdani fuzzy inference system (FIS) and earlier intuitionistic fuzzy set (IFS) based FIS in terms of RMSE. Reduced amount of RMSE in predicting score of positivity confirms outperformance of proposed IFIS over Mamdani FIS and earlier proposed IFS based FIS.