社交关系挖掘研究综述

发布时间：2019-02-13 13:23

【摘要】：随着在线社交应用和媒体的迅速扩散,在线社交网络(Online Social Network,OSN)已将我们的日常生活与网络信息空间连接起来.这些连接产生了大量的数据,不仅包括传播信息,还包括用户行为.社交关系挖掘的研究是社交网络挖掘中的一个重要领域,为我们对网络的形成机理、用户的交互模式和动态机制的理解提供了一个机会.社交关系(Social Ties)是社交网络中人与人连接和交互的纽带,也是社交网络中信息传播的基础.从计算学的观点来看,社交关系挖掘的研究包括社交关系的形成机理、社交关系的语义化以及基于社交关系人与人之间的交互.该文综述性地分析了这3个方面的研究现状,具体来说,在社交关系形成机理方面介绍关系链接预测,在基于社交关系的交互方面介绍关系交互预测,在社交关系语义化方面介绍关系类型预测.首先给出社交网络分析问题的形式化描述和相关概念、常用数据,然后分别介绍关系链接预测、关系类型预测和关系交互预测3个方面的方法、理论和模型,并给出重要的应用实例及其效果.最后,该文给出了未来工作的展望.
[Abstract]:With the rapid spread of online social applications and media, the online social network (Online Social Network,OSN) has connected our daily lives to the cyberspace. These connections produce a great deal of data, not only spreading information, but also user behavior. The research of social relationship mining is an important field in social network mining, which provides an opportunity for us to understand the formation mechanism of network, user interaction mode and dynamic mechanism. Social relationship (Social Ties) is not only the link between people and people in social network, but also the basis of information transmission in social network. From a computational point of view, the research of social relationship mining includes the formation mechanism of social relationship, the semantic of social relationship and the interaction between people based on social relationship. In this paper, the research status of these three aspects is summarized and analyzed. Specifically, the relationship link prediction is introduced in the formation mechanism of social relations, and the interaction prediction in interaction based on social relations is introduced. This paper introduces the prediction of the relationship types in the aspect of the semantics of social relations. Firstly, the formal description and related concepts of social network analysis problems are given, and the commonly used data are given. Then, the methods, theories and models of relationship link prediction, relationship type prediction and relationship interaction prediction are introduced, respectively. An important application example and its effect are given. Finally, the paper gives the prospect of future work.
【作者单位】：安徽大学计算智能与信号处理教育部重点实验室;安徽大学信息保障技术协同创新中心;清华大学计算机科学与技术系;
【基金】：国家“八六三”高技术研究发展计划项目基金(2015AA124102) 国家社会科学重大基金(13&ZD190) 国家自然科学基金(61402006,61175046) 安徽省高等学校省级自然科学研究基金(KJ2013A016) 安徽省自然科学基金(1508085MF113) 教育部留学回国人员科研启动基金(第49批) 安徽大学高层次人才需求计划项目资助~~
【分类号】：TP391.3;TP393.09

【相似文献】