一种基于排名聚合的社交网络关键用户挖掘方法.docxVIP

一种基于排名聚合的社交网络关键用户挖掘方法 梁耀洲　郭强　刘建国 摘 要：带有时间属性的动态社交网络逐步成为社交网络研究热点。相比静态网络，动态网络考虑用户交互发生的先后顺序，能够更直接地描述用户的交互关系和顺序。传统社交网络挖掘方法往往从用户交互路径进行评估，忽略了动态社交网络中交互时间片段的相互影响。综合考虑用户的交互顺序与时间影响，采用超邻接矩阵描述动态网络，并用排名聚合理论对用户影响力进行综合排名，提出了一种基于排名聚合的社交网络关键用户识别方法。Workspace实证数据集显示，该方法在准确率对比结果中，Spearman相关系数最多提高了13.45%，说明该方法在社交网络关键用户挖掘中具有适用性和有效性。 关键词：社交网络;超邻接矩阵;排名聚合;关键用户挖掘 DOI：10. 11907/rjdk. 192562 中图分类号：TP391 文献标识码：A 文章编号：1672-7800（2020）003-0186-04 A Method of Key User Identification for Social Network Based on Ranking Aggregation LIANG Yao-zhou1， GUO Qiang1， LIU Jian-guo2 （1. Research Center of Complex Systems Science， University of Shanghai for Science and Technology， Shanghai 20XX93， China; 2. Institute of Fintech， Shanghai University of Finance and Economics， Shanghai 20XX33， China） Abstract：In recent decades， the dynamic social network with time property has become a hot topic of social network research. Compared with static network， dynamic network considers the sequence of user interaction and can describe the interaction relationship and sequence of users more directly. Traditional social network mining methods are often evaluated by the communication path of users and ignore the interaction time segments in dynamic social networks. In this paper， considering the influence of the interaction sequence and time of users， the hyper adjacency matrix is used to describe the dynamic network， and the ranking aggregation theory is used to rank users comprehensively， and a ranking aggregation based key user identification method of social network is proposed. The result of Workspace data set shows that compared with the traditional method， Spearman correlation coefficient of this method is up to 13% higher in the result of accuracy comparison， indicating the applicability and effectiveness of this method in social network key user mining. Key Words： social network; super adjacency matrix; ranking aggregation; key user identification 0 引言 動