摘要
随着信息技术快速深入发展,为读者提供高效、差异化的服务逐渐成为现代图书馆提高服务职能的重要内容。随着图书馆读者和图书数量的迅速增加,读者-图书评分矩阵稀疏性问题日趋显著。为解决此问题,提出了一种提升图书推荐精度的协同过滤改进算法,根据读者-图书评分矩阵,建立LDA模型,得到读者-图书概率矩阵,根据属性对图书进行聚类,并对读者-图书概率矩阵进行裁切,然后,结合上下文信息,引入时间因子改进传统协同过滤算法相似度计算公式。实验表明,提出了协同过滤改进算法,能有效提升图书推荐效果。
With the rapid and in-depth development of information technology, providing efficient and differentiated services for readers gradually becomes an important content of the modern library to improve the service function. With the fast increasing of the readers and books, the problem of the sparsity of the reader-book rating matrix becomes more and more obvious. To solve the problem, this paper proposed an improved collaborative filtering algorithm for increasing book recommendation accu- racy, built an LDA model according to the reader-book rating matrix and got reader-book selection probability matrix, and then, we clustered the book set by book properties, and cut matrix by cluster results. Finally, we introduced time factor ac- cording to traditional collaborative filtering algorithm to improved similarity calculation formula. Experimental results show that the proposed algorithm can effectively improve the effect of book recommendation.
作者
柯秀文
KE Xiu-wen(Software College of Shangqiu Polytechnic,Shangqiu 476001)
出处
《微型电脑应用》
2018年第8期18-20,共3页
Microcomputer Applications
基金
河南省科技厅基金项目(132300410445)
商丘市社科规划基金项目(SKG-2015-211)
关键词
协同过滤
推荐算法
读者
图书
Collaborative filtering
Recommendation algorithm
Readers
Books