基于用户兴趣主题模型的个性化推荐研究  

Personalized Recommendation Research Based on User Interest Topic Model

作　　者：熊回香[1] 杨雪萍[1] 高连花 Xiong Huixiang Yang Xueping Gao Lianhua(Department of Information Management, Central China Normal University, Wuhan 430079 Best Buy Technology Co., Ltd, Shenzhen 518067)

机构地区：[1]华中师范大学信息管理学院,武汉430079 [2]优购科技有限公司,深圳518067

出　　处：《情报学报》2017年第9期916-929,共14页Journal of the China Society for Scientific and Technical Information

基　　金：国家社会科学基金项目"大众分类中标签间语义关系挖掘研究"(12BTQ038)

摘　　要：Web2.0的发展带来网络信息指数增长,给用户带来大量的信息和知识的同时,也引发了"信息过载"和"信息迷航"等问题,在一定程度上削弱了用户检索体验,而以"用户为中心"的个性化信息服务成为学术界和产业界关注的热点。已有研究大部分建立的是单用户兴趣模型,忽略了某些系统为群体用户提供个性化服务的需要等,本文在对社会化标签用户特征进行关联分析后,构建了基于社会化标签的单用户兴趣模型和群用户兴趣模型,并在此基础上,借鉴协同过滤算法的思想,架构了基于标签的单用户和群用户个性化信息服务流程框架,以期为社会化标注系统中信息服务的优化和实施提供理论与应用参考。In a multi-media platform of a network, apart from social websites that allow users to release and add tags, more and more information resource sharing systems are open to all user permissions in organizing and managing items and tags. This paper analyzed the advantages and disadvantages in applying recommendation technology to solve the difficulty in obtaining useful information. This paper constructed a personalized information recommenda- tion model based on the folksonomy system. It proposes the design of three recommendation components based on the three dimensions of resource, tag, and user, respectively, and combines the three candidate sets to realize a user interest prediction. The paper applies Douban Reading as an example to describe the proposed model and displays the results of personalized information.

关 键 词：大众分类 社会化标签 用户兴趣模型 个性化推荐 

分 类 号：G252[文化科学—图书馆学]