一种融合近邻用户影响力的矩阵分解推荐算法

Matrix Factorization Recommendation Algorithm Based on the Influence of Nearest Neighbor Users

随着电子商务的高速发展,推荐系统越来越受到各大电商的重视.近些年来,随着矩阵分解算法应用于推荐系统,使得其推荐精度得到了极大地提高.但传统的基于矩阵分解的推荐算法忽视了用户间存在互相影响这一事实,导致其推荐精度受到一定影响.基于此问题本文提出了一种近邻用户影响力的数学模型,综合考虑了近邻用户对于目标用户的影响.针对用户间影响力的复杂性,本文利用Cloud-model模型从统计学意义上考虑用户评分的相似性并以此获取目标用户的近邻,进而将目标用户近邻的影响力整合到矩阵分解算法中.经实验表明此算法有效提高了推荐算法的精度.

With the rapid development of e-commerce,recommendation systems are becoming more and more important. In recentyears, with the application of matrix factodzation algorithm to recommendation systems, the recommendation accuracy has been greatlyimproved. But the traditional algorithm based on matrix factorization ignores the fact that there is mutual influence among users, whichmakes the recommendation accuracy affected. Based on this problem,this paper presents a mathematical model of the influence of thenearest neighbor users, taking into account the influence of the nearest neighbor users on the target users. According to complexity ofuser influence, this paper uses the Cloud-model model to Consider the similarity of user ratings from the statistical significance and toobtain the target user neighbors, which will influence the integration to the target user nearest neighbor matrix factorization algorithm.Experimental results show that the proposed algorithm improves the accuracy of the proposed algodthra.