Exploring new frontiers: Novelty in recommender systems

Abstract

With the unimaginable vast amount of data on the World Wide Web, finding useful information has become increasingly challenging. To address this information overload, recommender systems have been developed. This paper presents the collaborative filtering approach, and a new novelty detection approach so to improve the ranking of novel items. These personalised recommender system keeps the user’s interest by recommending items based on his interests and the order of occurrences. The automatic detection of novel items leads to enhanced experience that adds more information to already known information by user. The past behaviours of the user assist the user in making more effective decisions that enhances the satisfaction level of the user.