Enhancing collaborative filtering by frequent usage patterns

Recommender systems contribute to the personalization of resources on the Web sites and information retrieval systems. In this paper, we present a hybrid recommender system using a user based approach which combines predictions based on Web usage patterns and rating data. We suggest a new technique that takes into account frequent patterns in order to compute correlations between users and select neighbors. Then, we combine this technique with collaborative filtering using Pearson correlation metric. The aim of this combination consists in the evaluation of the impact of each technique on recommendations. The performance of our system is tested without and by combining predictions in terms of accuracy and robustness. The different tests show that the more the navigational based technique is involved in the recommendation process, the more the best predictions are accurate and the system is robust.