DERIVING COLLECTIVE RECOMMENDATION WITH ASPECT-BASED SENTIMENT AND SOCIAL INFLUENCE

Despite the vast amount of restaurant information shared on social platforms, users often face difficulties identifying suitable options efficiently. Ratings are constrained by information narrowness, while textual reviews pose challenges due to information overload. Furthermore, restaurant recommendations based solely on ratings lack objectivity, as individual preferences differ. Users cannot judge whether a restaurant is worth visiting without credible information. However, few existing studies integrate semantic analysis with multidimensional orientation and social influence to make recommendations, leaving a gap in objective and comprehensive analysis. This research proposes a synthesis collective recommendation approach utilizing machine learning with aspect-based sentiment and social influence analyses. The proposed approach can appropriately adjust ratings as a basis for deciding the list of recommendations, considering location and preference factors. Experimental results show that the proposed mechanism significantly enhances users’ ability to find restaurants that meet their needs, thereby improving business opportunities.