基于用户—资源关联的社会化小众推荐模型研究  

Research on the Social Minority Recommendation Model Based on User- Resources Association

作　　者：胡吉明[1] 张蔓蒂[2] 

机构地区：[1]武汉大学信息资源研究中心,湖北武汉430072 [2]华为技术有限公司武汉办事处,湖北武汉430010

出　　处：《情报理论与实践》2014年第4期123-126,118,共5页Information Studies:Theory & Application

基　　金：国家自然科学青年基金项目"社会网络环境下基于用户-资源关联的信息推荐研究"(项目编号:71303178);中国博士后科学基金特别资助项目"社会网络环境下基于用户关系的信息推荐研究"(项目编号:2013T60750)的成果

摘　　要：社会网络环境下用户小众化需求和资源长尾分布特征,要求推荐策略从大众化向小众化转变。文章在分析社会网络环境下用户小众化需求特征和推荐策略变革的基础上,引入用户与资源之间的关联关系,提出了社会化小众推荐策略。根据物理动力学中物质扩散和热传导能量分配理论,分析了基于用户—资源二部关联图的社会化小众推荐机理,构建了社会化小众推荐模型,并在典型的社会化书签服务中进行了相应的实证分析,以期提升资源推荐的个性化效果。The characteristics of user minority needs and long-tail resources distribution in the social network environment re- quire that recommendation strategies should transform from popularity to minority. Based on the analysis of the characteristics of user minority needs and the transformation of recommendation strategies in the social network environment, this article brings in the asso- ciated relations between users and resources, and proposes the social minority recommendation strategies. Furthermore, according to the energy distribution theory of substance diffusion and heat conduction in physical dynamics, the article analyzes the social minori- ty recommendation mechanism based on user-resources bipartite association diagram, and constructs a social minority recommenda- tion model. Finally, an empirical study is conducted in a typical social bookmarking service so as to improve the personalized effects of resources recommendation.

关 键 词：社会网络 社会化推荐 二部关联图 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]