The Role of XAI in User-Centric Recommender Systems Using Collaborative and Content-Based Approaches

Recommendation systems are integral to new-age digital platforms, influencing user behavior by offering personalized suggestions. However, their lack of opacity often sparks concerns about user autonomy and fairness, as users are left unenlightened of how recommendations are generated. This paper addresses the importance of explainable AI (XAI) in recommendation systems to build user trust. We implemented collaborative and content-based movie recommender systems using the MovieLens 100k dataset, and extended the content-based approach to the TMDB 5000 dataset demonstrating how explainability can be integrated into recommendation models to provide users with insights into the reasoning behind specific movie suggestions. By generating simple yet effective explanations based on movie genre, user ratings, commonality among users, similarity score, keywords, and other relevant features, our approach intends to help users understand the reasoning behind their recommendations. Our comparative analysis highlights the advantages of explainable recommendation systems over non-explainable ones, showing that the inclusion of explanations significantly improves user satisfaction by reducing confusion and increasing transparency.