基于情境感知的用户个性化兴趣建模  

Personalized user interest modeling based on context aware

作　　者：葛桂丽 袁凌云[1] 王兴超[2] 

机构地区：[1]云南师范大学信息学院,昆明650500 [2]云南大学信息学院,昆明650031

出　　处：《计算机应用研究》2017年第4期995-999,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61561055);国家教育部人文社会科学青年基金资助项目(13YJCZH233);云南省教育厅一般资助项目(2015Y001);云南省应用基础研究计划资助项目(2013);云南师范大学研究生科研创新基金资助项目(yjs201534)

摘　　要：用户兴趣模型作为个性化信息推送的基础和核心,其性能的好坏直接关系到个性化推送服务的质量,如何全面了解用户动态的兴趣需求并及时为用户提供其实时有效的兴趣,是目前用户兴趣模型研究的热点问题。主要利用情境感知、用户行为等多维因素建立一种M-C-W用户兴趣模型,提出结合用户显式和隐式兴趣度的计算,挖掘出用户显式兴趣和隐式兴趣的关联性,实现多维动态情境兴趣的结合。通过多角度综合计算用户的兴趣度,实时而准确地表达用户的兴趣爱好。最后,通过实验验证了该模型的有效性和可靠性。As the basis and core of the personalized recommendation, the quality of the user interest model was of close concern to the service quality of personalized recommendation. How to fully understand the dynamic user interest and provide timely effective interest information to users, is a hot issue in the current user interest model. Based on the context aware technology, this paper proposed a new multiple-context-weight（M-C-W） user interest model and a updating processed mechanism. The M-C-W model was based on multi-dimensional factors such as context awareness and user behavior. The user＇ s real-time, complete as well as accurate interests were expressed by the dynamic weighted fusion method, which was using the combination of explicit and implicit interest. The experimental results prove that the M-C-W model is feasible and reliable.

关 键 词：用户兴趣模型 情境感知 个性化信息推送 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]