A Hybrid Collaborative Filtering Recommendation Algorithm Using Double Neighbor Selection

Abstract

The traditional collaborative filtering algorithms are more successful used for personalized recommendation. However, the traditional collaborative filtering algorithm usually has issues such as low recommendation accuracy and cold start. Aiming at addressing the above problems, a hybrid collaborative filtering algorithm using double neighbor selection is proposed. Firstly, according to the results of user’s dynamic similarity calculation, the similar interest sets of the target users may be dynamically selected. Analyzing the dynamic similar interest set of the target user, we can divide the users into two categories, one is an active user, and the other is a non-active user. For the active user, by calculating the trust degree of the users with similar interests, we can select the user with the trust degree of TOP-N, and recommend the target user. For the non-active user, the neighbor user may be found according to the similarity of the user on some attributes, and them with high similarity will be recommend to the target user. The experimental results show that the algorithm not only improves the recommending accuracy of the recommendation system, but also effectively solves the problem of data sparseness and user cold start.