情報処理学会 第86回全国大会 会期:2024年3月15日~17日

4M-05
A Group-based Approach to Enhance Diversity and User Engagement in Movie Recommender Systems
○任 宝峰,木實新一(九大)
Current movie recommender systems often prioritize accuracy, overlooking the diversity of recommendations and user interaction experience. Addressing these gaps, we introduce a system featuring rating lists from user groups alongside user-controllable options. Besides preference-based movie recommendations, our system aggregates rating lists from user groups that share common characteristics with the user. Recognizing that users have varied tastes, including low-rated movies in the rating list and exposing rating lists given by user groups can offer valuable and diverse reference than user preference-based approach. Additionally, our system incorporates control options, aiming to improve the interaction experience by allowing users to participate more actively in the recommendation process.