基于用户—资源—词汇三部图的社会化推荐算法设计与实现  

Design and Implementation of Social Recommendation Algorithm Based on User-Object-Topic Tripartite

作　　者：胡吉明[1] 林鑫[1] 

机构地区：[1]武汉大学信息资源研究中心,湖北武汉430072

出　　处：《情报理论与实践》2016年第3期130-134,共5页Information Studies:Theory & Application

基　　金：国家自然科学基金青年项目"社会网络环境下基于用户-资源关联的信息推荐研究"的成果;项目编号:71303178

摘　　要：面向用户小众化需求和长尾资源特征的社会化推荐实现对于提升网络服务质量具有重要作用。文章根据物理动力学中物质扩散和热传导能量分配机理,并将词汇引入关联图中,探讨了基于用户—资源—词汇三部图的推荐实现机理;分别进行了用户—资源二部图和资源—词汇二部图中基于能量分配加权的推荐算法实现研究,提出了二部图和能量分配双重加权的三部图推荐实现策略。最后,通过大量的准确性和多样性的实验,验证了所提策略和方法的有效性及优越性。The realization of social recommendation for user minority needs and long tail resources features plays an important role in improving the quality of network service. According to the mechanism of energy distribution based on mass diffusion and heat spreading in physic kinetics, this paper introduces topics into tripartite to discuss the realization mechanism of recommendation based on user-object-topic tripartite. Then, the paper analyzes the realization of recommendation algorithm based on weighted energy distribution in user-object-topic tripartite and object-topic bipartite and proposes the recommendation strategy of bipartite and doubleweighted energy distribution tripartite. Finally, the paper verifies the effectiveness and superiority of the proposed strategy and method through a large amount of experiments.

关 键 词：三部关联图 社会化推荐 算法 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术] G350[自动化与计算机技术—计算机科学与技术]