基于用户行为综合分析的微博用户影响力评价方法  

Method of evaluating micro-blog users' influence based on comprehensive analysis of user behavior

作　　者：齐超[1] 陈鸿昶[1] 于洪涛[1] 

机构地区：[1]国家数字交换系统工程技术研究中心,郑州450002

出　　处：《计算机应用研究》2014年第7期2004-2007,共4页Application Research of Computers

基　　金：国家"863"计划资助项目(2011AA010603)

摘　　要：通过对用户转发、评论和提及三种行为的综合分析,运用统计分析方法就微博用户的不同行为对传播影响力的贡献进行度量,提出一种基于行为权值分配的PageRank算法,对传播影响力大小进行定量的分析。采用微博数据的实验结果表明,该算法比其他传统算法对传播影响力判断的准确性更高,且用户特征中最能反映用户传播影响力大小的是用户转发情况。Through the comprehensive analysis of retweet,comment,mention and statistical analysis of different contribution of different behaviors for influence,this paper proposed a PageRank algorithm based on user behavior weight assigned to evaluate spreading influence quantitatively. Experimental results based on micro-blog data show that this algorithm is more accurate and effective than other traditional methods. And the most effective characteristic of user in evaluating spreading influence is the situation of user retweet.

关 键 词：微博用户 传播影响力 用户行为 PAGERANK 影响覆盖率 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]