Integrating Fuzzy Logic into Lexicon-Based Sentiment Analysis: A Hybrid Framework

Abstract

Sentiment analysis, a fundamental natural language processing (NLP) task, aims to classify textual data into positive, negative, or neutral sentiments. Traditional lexicon-based methods rely on predefined sentiment dictionaries to assign polarity scores, but they often struggle with ambiguous, mixed, or context-dependent sentiments. To address this limitation, fuzzy logic-integrated approach has been proposed here for sentiment classification. Fuzzy logic offers a flexible and human-like reasoning approach capable of handling uncertainty and partial truths, making it well-suited for nuanced sentiment assessment. This paper presents a hybrid model that utilizes fuzzy logic to polish sentiment classification. Experimental results demonstrate that combining fuzzy logic with lexicon-driven method can significantly improve classification accuracy.