摘要
传统的协同过滤相似性度量方法主要考虑用户评分之间的相似性,缺少对评分差异性的考虑。文中将用户评分关系分为差异部分和相关部分,提出了一种基于用户评分差异性和相关性的相似性度量方法。该方法在非极其稀疏数据集下有较好的推荐效果。针对该方法在稀疏数据集下存在推荐不准确的问题,采用预填充方法对其进行改进。实验表明,该方法在预填充后的推荐精度得到明显提高。
The traditional similarity measurement in collaborative filtering mainly pays attention to the similarity between users＇ ratings,lacking the consideration of difference of users＇ ratings.This paper divided the relationship of users＇ ratings into differential part and correlated part,and proposed a similarity measurement based on the difference and the correlation of users＇ ratings on the non-sparse dataset.In order to solve the problem that the algorithm＇s recommendation is not accurate in spare dataset,this paper improved this algorithm by prefilling the vacancy of rating matrix.Experiment results show that this algorithm can significantly improve the accuracy of recommendation after prefilling the rating matrix.
作者
王劲松
蔡朝晖
李永凯
刘树波
WANG Jing-song;CAI Zhao-hui;LI Yong-kai;LIU Shu-bo(School of Computer,Wuhan University, Wuhan 430072, Chin)
出处
《计算机科学》
CSCD
北大核心
2018年第5期190-195,共6页
Computer Science
基金
国家自然科学基金(41671443)资助
关键词
协同过滤推荐
差异性
相关性
预填充
Collaborative filtering recommendation
Difference
Correlation
Prefilling
