首页 | 官方网站   微博 | 高级检索  
     

一种基于协作过滤的电子图书推荐系统
引用本文:曾庆辉,邱玉辉.一种基于协作过滤的电子图书推荐系统[J].计算机科学,2005,32(6):147-150.
作者姓名:曾庆辉  邱玉辉
作者单位:西南师范大学计算机与信息科学学院,重庆,400715
基金项目:本文研究得到重庆市自然科学基金资助,项目编号:CSTC,2004BB2086.
摘    要:推荐系统中最常见信息过滤技术是基于内容的过滤和协作过滤,协作过滤由于有其自身的优点得到迅速发展,并得到广泛应用,但传统的协作过滤算法存在着稀疏性、扩展性和同义性等问题。本文提出一种基于评价矩阵列向量的图书协作过滤算法,并把这个算法应用到了一个数字图书馆的电子图书推荐系统中。此图书协作过滤算法主要计算图书之间的相似度而不是用户之间的相似度,可以大大降低计算量。实验也表明,这个算法比传统的基于用户的协作过滤算法有优势。

关 键 词:推荐系统  协作过滤  图书推荐  评价矩阵  数字图书馆

An E-Book Recommender System with Collaborative Filtering
ZENG Qing-hui,QIU Yu-Hui.An E-Book Recommender System with Collaborative Filtering[J].Computer Science,2005,32(6):147-150.
Authors:ZENG Qing-hui  QIU Yu-Hui
Affiliation:ZENG Qing-Hui,QIU Yu-Hui School of Computer and Information Science,Southwest China Normal University,Chongqing 400715
Abstract:Collaborative filtering and content-based filtering are the most common information filtering technology in recommender system. Collaborative filtering is becoming the popular one and has been used widely because of its good quality. But traditional collaborative filtering algorithm has the shortcomings of sparsity,scalability and synonymy. In this paper,we present a new collaborative filtering algorithm base on the column-vector of the evaluations matrix for an e-book recommender system in the digital library. The algorithm computes the similarity of books instead of the shailarity of users,which can remarkably alleviate the workload. Our experiments suggest that the algorithm provides better performance than user-based algorithm.
Keywords:Recommender system  Collaborative filtering  Book recommending  Evaluations matrix  Digital library  
本文献已被 CNKI 维普 万方数据 等数据库收录!
点击此处可从《计算机科学》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号