A New Weighted Similarity Method Based on Neighborhood User Contributions for Collaborative Filtering
Xuefeng Zang, Tianqi Liu, Shuyu Qiao, Wen-Zhu Gao, Jiatong Wang, Xiaoxin Sun, Bangzuo Zhang · 2016
Collaborative filtering is the most successful and widely used method in personalized recommendation service since it is simple and effective. The key point is to find similar users or items through the user-item rating matrix. However, traditional collaborative filtering do not consider the information about items rated by neighbors and that not rated by the target user, and there has another important factor that preference of each user is different. This paper proposes a new similarity measuring method that takes into account the proportion of co-rating, the user rating preference and the different contributions of other users to the target user. The method is implemented and compared with many other state-of-the-art similarity measures in two real data sets, MovieLens-100K and FilmTrust. The experimental results show that the proposed method most often outperforms the traditional collaborative filtering measures.