基于K-means的语义协同过滤推荐算法  

Research on semantic collaborative filtering recommendation algorithm based on K-means

作　　者：印国成 YIN Guocheng(School of Information Engineering,Yangzhou University,Yangzhou 225127,China;Guangling College of Yangzhou University,Yangzhou 225009,China)

机构地区：[1]扬州大学信息工程学院,江苏扬州225127 [2]扬州大学广陵学院,江苏扬州225009

出　　处：《扬州大学学报（自然科学版）》2018年第1期46-49,共4页Journal of Yangzhou University：Natural Science Edition

基　　金：江苏省高校自然科学基金资助项目(14KJB520041);江苏省科技厅产学研前瞻性资助项目(BY20150 61-06;BY20150 61-08)

摘　　要：传统的基于图书和读者的协同过滤方法缺乏语义知识,易混杂不符合读者喜好的噪音数据,从而影响聚类效果和推荐的准确度.针对该问题,提出一种基于K-means的语义协同过滤推荐算法.为了反映读者对图书的偏爱程度,首先定义读者-图书关联矩阵,然后通过K-means聚类算法寻找相邻集合,在聚类过程中兼顾关联矩阵和语义知识,分别计算读者和图书的相似度,最后通过相似程度排序向用户推荐图书.结果表明,该算法在保证计算效率的情况下能显著提高推荐的准确度.The author considers the condition that the traditional user-based or the item-based collaborative filtering algorithms lack semantic knowledge,easily confound noise data which are not in line with users’preference,and affect the clustering effect and recommendation accuracy.In order to solve the problem,a semantic collaborative filtering recommendation algorithm based on K-means is proposed.To express the degree of preference for the readers of the books,the author defines the reader-book association matrix,and then uses the K-means clustering algorithm to find the neighbor set which can also take semantic knowledge into consideration during clustering procedures and can improve recommendation accuracy.Experimental results show that the algorithm presented in this paper significantly outperforms than others without reducing computational efficiency.

关 键 词：推荐 协同过滤 语义 K-MEANS算法 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术] G254.924[自动化与计算机技术—计算机科学与技术]