Deep Learning-Based Multi-Criteria Recommender System: Leveraging Autoencoders for Improved Personalization

Multi-Criteria Recommender Systems (MCRSs) are a promising method that takes into account several elements of the user's preferences in order to improve recommendation accuracy. However, most existing research has focused on single-criterion ratings, overlooking the potential of multi-criteria considerations in deep learning-based recommender systems. This paper introduces a novel deep learning-based algorithm for MCRSs, leveraging deep autoencoders to uncover complex user-item interactions based on multi-criteria preferences. The proposed algorithm, evaluated using the TripAdvisor multi-criteria dataset, demonstrates superior recommendation capabilities compared to traditional methods, achieving a significantly lower RMSE of 0.975. The results highlight the algorithm's effectiveness in providing accurate and personalized recommendations tailored to diverse user preferences, marking a significant advancement in MCRS performance.