Collaborative Filtering Recommendation Algorithm Based on Multi-relationship Social Network

Abstract

For solving the data sparsity of traditional algorithms, this paper proposes a novel collaborative filtering recommendation algorithm based on multi-relationship social network. On the basis of traditional matrix decomposition model, the proposed algorithm obtains the trust and trust feature matrix by integrating the user preferences of multi-relationship social network, and then, the rating of the commodity is predicted by the social feature matrix, the commodity feature matrix, and the similarity of user rating preference. In order to verify the reliability of the proposed algorithm, the Epinions dataset is used to compare the performance of the algorithm with that of the existing social network recommendation algorithms. According to the experimental results, the proposed algorithm had smaller mean absolute error (MAE) and root mean square error (RMSE), indicating that it has effectively reduced the impact of data sparsity on recommendation results and improved the recommendation accuracy.