THE COMPARISON OF SEVERAL METHODS OF PROCESSING NO-RATED ITEMS IN COLLABORATIVE FILTERING ALGORITHM

摘要：

Collaborative Filtering is a very important technology in E-commerce. Unfortunately, with the increase of users and commodities, the user rating data is extremely sparse, which leads to the low efficient Collaborative Filtering recommendation system. To address these issues, many methods of processing no-rated items in Collaborative Filtering recommendation algorithm have been proposed, including algorithm without taking the no-rated items into account, and algorithms setting the ratings of no-rated items’ value as 0, half of the full score, average of the target item’s rating score, or the average of the target user’s rating score. This paper compares the several methods, and the experimental results show that the method of set ting no-rated items’ value as 0 is the best method in these methods.

关键词： Personalized Recommendation Userbased Collaborative Filtering Item-based Collaborative Filtering item similarity user similarity MAE

作者: Zhang Jinbo Lin Zhiqing Xiao Bo Zhang Chuang

作者单位: Pattern Recognition & Intelligent System Lab,Beijing University of Posts and Telecommunications, Beijing

会议类型: 国际会议

会议名称: 2009 2nd IEEE International Conference on Broadband Network & Multimedia Technology(2009年宽带网络与多媒体国际会议 IEEE IC-BNMT2009)

会议地点: 北京

会议语种:英文

页码: 873-876

在线出版日期: 2009-10-18（万方平台首次上网日期，不代表论文的发表时间）