Evolutionary heterogeneous clustering for rating prediction based on user collaborative filtering

Recommender systems play an important role in our life, which would help users to find what they are interested in. Collaborative filtering is the most widely used and successful method for personalized recommendation. In this paper, a novel heterogeneous evolutionary clustering is presented. The goal of our algorithm is to gather users with similar interest into the same cluster and to help users find items that fit their personal tastes best. The suggestions from friends with similar interest may be adopted with high probability. Firstly, items and users are regarded as heterogeneous individuals in the network. According to the constructed network model, states of individuals evolve over time. Individuals with higher scores would cluster into together and individuals with lower scores would get away. After many iterations, states of items and users would be stable. In light of stable states of heterogeneous individuals, they are clustered into several groups. Secondly, user-based collaborative filtering are adopted in each cluster. Similarities between individuals only in same cluster are computed not for all individuals in system. The target rating is calculated according to user-based collaborative filtering in its cluster. Diverse simulations show the efficiency of our proposed methods. Moreover, the presented method gains better prediction results than two existing preferable algorithms.