On the Effectiveness of Prioritized User-Profile and Detecting Active Users in Collaborative Filtering Recommender Systems

Recommender systems intend to help users find their interested items from among a large number of items. These systems especially those based on collaborative filtering methods have shown success on the Web. This paper emphasizes on the effectiveness of the prioritized user-profile and detecting active users as two basic approaches that could improve the quality of collaborative filtering recommenders. The first approach is based on [1] and tries to implement more personalized recommendation by assigning different priority importance to each of the features of the user-profile for different users. The second approach tries to reduce the effort needed to find similar users and at the same time increase the quality of recommendations by using the opinions of active users in the system. These two techniques are compared with the standard user-based Pearson algorithm (Breese et al., 1998) on book and movie datasets.