Improving sentiment analysis using preprocessing techniques and lexical patterns. (23rd September 2021)

• Record Type: Journal Article
• Authors: Cagnoni, Stefano; Ferrari, Laura; Fornacciari, Paolo; Mordonini, Monica; Sani, Laura; Tomaiuolo, Michele
• Is Part Of: International journal of data analysis techniques and strategies. Volume 13:Number 3(2021)
• Page Start: 171
• Page End: 185
• Publication Date: 2021-09-23
• Subjects: sentiment analysis -- natural language processing -- POS tagging -- feature weighting -- word stemming -- bag-of-words representation -- tf-idf -- Penn Treebank Tagset -- support vector machines -- naïve Bayes multinomial classifier; Electronic data processing -- Periodicals; Database searching -- Periodicals; 005
• Languages: English
• ISSNs: 1755-8050

Abstract

Sentiment analysis has recently gained considerable attention, since the classification of the emotional content of a text (online reviews, blog messages etc.) may have a relevant impact on market research, political science and many other fields. In this paper, we focus on the importance of the text preprocessing phase, proposing a new technique we termed lexical pattern-based feature weighting (LPFW) that allows one to improve sentence-level sentiment analysis by increasing the relevance of the features contained in particular lexical patterns. This approach has been evaluated on two sentiment classification datasets. We show that a systematic optimisation of the preprocessing filters is important for obtaining good classification accuracy. Also, we show that LPFW is effective in different application domains and with different training set sizes.