A Modified Fuzzy Sentiment Analysis Approach Based on User Ranking Suitable for Online Social Networks

摘要

Processing Sentiment Analysis (SA) in social networks has lead decision makers to value opinion leaders who can sway people's impressions concerning certain business or commodity. Yet, a tremendous scarceness of considering perspectivism, while computing text polarity, has been spotted. Considering perspectivism in SA can help in the production of polarity scores that represent the perceptible sentiment within the content. This emphasizes the necessity for integrating social behavior (user influence factor) with SA (text polarity scores), providing a more pragmatic portrayal of how text-recipients comprehend the message. In this paper, a novel model is proposed to intensify SA process in Twitter. In the achievement of such, UCINET tool and Artificial Neural Networks (ANN) are used for social network analysis (SNA) and users ranking respectively. For sentiment classification, a hybrid approach is presented --lexicon-based technique (TextBlob) along with fuzzy classification technique --to handle language vagueness as well as for an inclusive analysis of tweets into seven classes; for the purpose of enhancing final results. The proposed model is practiced on data collected from Twitter. Results show a significant enhancement in the final polarity scores, associated with the analyzed tweets, representing more realistic sentiments.