Aspect-based Sentiment Analysis in Microblogs through Fuzzy Logic Techniques

The advent of social networks has elevated user-generated microblog text to a valuable asset for sentiment analysis, particularly among university students. Their microblogs serve as a portal to their psychological states and emotional inclinations. Traditional sentiment analysis methods, however, encounter limitations when grappling with the intricacies and nuanced variations in emotional expression. This study introduces and implements an innovative framework for aspect-based sentiment analysis (ABSA) of university students’ microblogs, integrating fuzzy logic techniques. A methodology is presented for the identification of emotional tendencies in microblog content, amalgamating sentiment analysis with intuitionistic fuzzy set (IFS) theory. This approach effectively addresses the inherent multidimensionality and ambiguity in emotional expressions. Further, an aspect-based sentiment classification model, rooted in the fuzzy neural network (FNN), has been developed. This model enhances both the precision and sensitivity of sentiment classification. Results from this study signify a marked enhancement in ABSA over conventional methods. This advancement offers substantial support for data-driven decision-making in areas such as higher education management and psychological health services, underscoring the practical applications of the research.