A Hybrid Multi-Criteria Collaborative Filtering Model for Effective Personalized Recommendations

Recommender systems act as decision support in supporting users selecting the right choice of items or services from a high number choices an overloaded search space. However, such have difficulty dealing with sparse rating data. One way to deal this issue is incorporate additional explicit information, also known side information. information requires some action and often not always available. Accordingly, study presents hybrid multi-criteria collaborative filtering model. The proposed model exploits ratings, implicit similarity, similarity transitivity global reputation concepts expand space potential recommenders. This expansion will enhance prediction accuracy coverage when applied data situations. To show effectiveness model, set experiments are conducted on two real-world datasets, Yahoo! Movies TripAdvisor. experimental results demonstrate superiority compared existing filtering-based recommendation methods under variety evaluation metrics.