基于领域知识的个性化推荐模型及其应用研究  

Domain Knowledge Based Personalized Recommendation and Its Application

作　　者：陈全[1,2] 张玲玲[1,2] 石勇[1,2] 

机构地区：[1]中国科学院研究生院管理学院 [2]中国科学院研究生院虚拟经济与数据科学研究中心

出　　处：《管理学报》2012年第10期1505-1509,共5页Chinese Journal of Management

基　　金：国家自然科学基金资助项目(71071151;70921061);中国科学院研究生院院长基金资助项目(085102HN00)

摘　　要：首先,讨论了传统电子商务领域个性化推荐算法及其应用,提出传统推荐方法通常忽视了领域知识,因而在实际应用中存在局限性;然后,讨论了领域知识的相关概念和分类,并在此基础上建立了结合领域知识的两阶段推荐模型;最后,基于网络超市数据进行了实证研究,实验结果表明改进后的推荐算法效果要明显优于传统的个性化推荐算法。This paper discusses the traditional personalized recommendation methods and their ap- plications in e commerce. Then, it argues that classic collaborative filtering usually does not take do- main knowledge into consideration and its limitation in real world application. It further discusses the concept of domain knowledge, and build a two-steps recommendation model which combines domain knowledge with traditional recommendation method. Finally, through empirical study, it is proved that personalized recommendation based on domain knowledge runs much better than traditional recommendation method in all evaluation indicators.

关 键 词：领域知识 个性化推荐 协同过滤 

分 类 号：C93[经济管理—管理学]