基于用户行为和结构洞的用户影响力评价模型  

User influence comprehensive evaluation model based on user behavior and structure hole principle

作　　者：魏建冬 覃锡忠[1] 贾振红[1] 牛红梅 WEI Jiandong;QIN Xizhong;JIA Zhenhong;NIU Hongmei(College of Information Science and Engineering,Xinjiang University,Urumqi 830046,China;Department of Network Monitoring,China Mobile Communications Group Xinjiang Limited,Urumqi 830046,China)

机构地区：[1]新疆大学信息科学与工程学院,新疆乌鲁木齐830046 [2]中国移动通信集团新疆有限公司网络监控部,新疆乌鲁木齐830046

出　　处：《现代电子技术》2019年第5期164-168,共5页Modern Electronics Technique

基　　金：教育部促进与美大地区科研合作与高层次人才培养资助项目(20101595);中国移动通信集团新疆有限公司研究发展基金项目(XTM2013-2788)~~

摘　　要：针对目前大多数用户影响力度量方法未能充分考虑非冗余信息的问题,结合结构洞原理,构建基于用户行为和结构洞的用户影响力评价模型。对用户行为进行量化,将其融入到网络边权中,并在此基础上对结构洞评价方法——网络约束系数进行改进,解决其未考虑用户行为对网络边权的影响,以及节点与邻近节点、次邻近节点三层拓扑关系的问题,最后构建用户影响力综合评价模型。通过新浪微博数据进行验证,排序实验与覆盖率实验结果验证了该模型用户影响力度量的准确性与有效性。Since the non-redundant information is not fully considered in current most user influence measurement methods,the user influence comprehensive evaluation model based on user behavior and structure hole principle is constructed.The user behavior is quantified,and integrated into the edge weight of the network.On this basis,the network constraint coefficients are improved by means of structural hole method to consider the influence of user behavior on the edge weight of the network,and solve the problem of three.layer topological relationship among the node,adjacent node and subjacent node.The user influence comprehensive evaluation model is constructed,and verified with the data from Sina MicroBlog.The results of sequencing experiment and coverage ratio experiment show that the user influence measurement of the model is accurate and valid.

关 键 词：用户影响力 非冗余信息 用户行为 结构洞 网络约束系数 覆盖率 

分 类 号：TN911.34[电子电信—通信与信息系统] TP393[电子电信—信息与通信工程]