Ontology-Based Recommender System for Bandung Region Eateries with Sentiment Analysis Enhancement

In recent years, finding high-quality services has become an increasingly complex challenge for users. Particularly in the fast-growing restaurant sector, consumers are faced with a huge number of options, making it difficult for them to choose a restaurant that suits their individual preferences and tastes. For example, in the Bandung area, there are many eateries that offer various types of food, drinks, and other facilities that can make it difficult for users to choose a place that suits their preferences and needs. Although there are many recommender systems available to help select a place to eat based on the user’s preferences, these systems often lack direct user involvement. To overcome the limitations that have been mentioned, this research aims to design an ontology-based conversational recommender system with sentiment analysis. This system offers ratings of dining places from real users as a result of sentiment analysis and personalized recommendations based on functional requirements for dining places in Bandung. The results generated by the system, based on both accuracy and user satisfaction metrics, revealed an impressive accuracy rate of 85,33%. Additionally, the study garnered positive feedback from users who utilized the system to find restaurants matching their preferences. The system’s high accuracy and positive user feedback show its effectiveness in meeting user needs and confirm its reliability as a tool for restaurant recommendations.