摘要
提出一种新的基于协同聚类的协同过滤矩阵分解方法,可以兼顾协同过滤算法的有效性和效率。首先将协同过滤中的用户项目评分矩阵分解为若干个潜在子矩阵。在此基础上,提出一种有效的矩阵分解算法,该算法能够充分利用每个聚类用户和项目元素之间的内在联系,并且因为不同簇之间的交互作用可以忽略不计,故在前一步得到的单个簇上可以同时执行矩阵分解算法,因而能提高运行效率。实验结果表明:利用矩阵分解的方法能够提高协同过滤的预测精度和效率。
This paper proposes a new collaborative filtering matrix decomposition method based on Collaborative clustering,which can take into account the effectiveness and efficiency of collaborative filtering algorithm.Firstly,the user item rating matrix in collaborative filtering is decomposed into several potential sub matrices.On this basis,an effective matrix decomposition algorithm is proposed.The algorithm can make full use of the internal relationship between each cluster user and project element,and because the interaction between different clusters can be ignored,the matrix decomposition algorithm can be executed simultaneously on a single cluster obtained in the previous step,so it can improve the operation efficiency.The experimental results show that the matrix decomposition method can improve the prediction accuracy and efficiency of collabo rative filtering.
作者
纪平
胡学友
杨文娟
刘学亮
JI Ping;HU Xue-you;YANG Wen-juan;LIU Xue-liang(School of Advanced Manufacturing Engineering,Hefei University,Hefei,230601;School of Computer and Information,Hefei University of Technology,Hefei,230009,China)
出处
《合肥学院学报（综合版）》
2020年第5期10-18,共9页
Journal of Hefei University：Comprehensive ED
基金
安徽省高校自然科学研究重点项目(KJ2018A0545)
安徽省高校学科(专业)拔尖人才学术资助项目(gxbjZD48)资助
关键词
协同过滤
矩阵分解
协同聚类
collaborative filtering
matrix factorization
co-clustering
