Item-Based Collaborative Filtering for Effective Board Game Recommendation

Abstract

Board game players often face difficulties in discovering suitable games, finding players, and locating places to play board games. This research presents a novel website addressing these issues by incorporating a board game recommendation system and a dedicated community platform for board game players. We developed and evaluated various recommendation models, including content-based filtering, user-based collaborative filtering, item-based collaborative filtering, and deep learning-based approaches. While the deep learning model exhibited the highest performance, the item-based collaborative filtering emerged as the most practical solution due to its effectiveness and implementation feasibility. Our experiments demonstrate the effectiveness of the item-based collaborative filtering model in recommending suitable board games to users. This website, equipped with this recommendation system and a community space, holds significant potential to enhance the overall board game experience for players.