基于BiasSVD和聚类用户最近邻的协同过滤算法研究  

Research on the Collaborative Filtering Algorithm Based on BiasSVD and Clustering User Nearest Neighbor

作　　者：李佳 张牧[1] LI Jia;ZHANG Mu(Chengdu Sport University,Chengdu 610041)

机构地区：[1]成都体育学院,成都610041

出　　处：《现代计算机》2020年第13期30-34,共5页Modern Computer

基　　金：四川省应用基础研究基金(No.2017JY0198);四川省哲学社会科学重点研究基地项目(No.TY2014201);四川网络文化研究中心项目(No.WLWH14-46)。

摘　　要：针对协同过滤推荐算法可扩展性问题,提出一种基于BiasSVD和聚类用户最近邻的协同过滤算法。通过BiasSVD对评分矩阵降维,使该算法能够扩展到大型数据集中。但考虑到其方式预测出的评分和实际评分有一定的误差,结合聚类用户最近邻实际评分与预测评分的平均差值来调整目标用户预测评分。实验表明,该算法能在一定程度上提高系统可扩展性。Regarding the problem of extendibility about collaborative filtering algorithms,this paper proposed a collaborative filtering algorithm based on BiasSVD and clustering user nearest neighbor.It reduced rating matrix dimensionality by BiasSVD,and the algorithm can be extended to a large data set.However,considering there have some certain error between the rating through its way and the actual one,the new algo⁃rithm combines with the average difference of the actual rating and forecast rating from clustering user nearest neighbor to adjust the target user prediction rating.Experimental results show that the proposed algorithm can improve the system extendibility to some extent.

关 键 词：协同过滤 BiasSVD 聚类用户最近邻 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]