一种改进PageRank的微博用户影响力计算方法  

Calculation Method of Microblog Users' Influence Based on Improved PageRank

作　　者：郑远飞 陈晓升[1] 王志文[1] 陈坚旋 陈珂[1] 

机构地区：[1]广东石油化工学院计算机与电子信息学院,广东茂名525000

出　　处：《广东石油化工学院学报》2016年第3期46-50,54,共6页Journal of Guangdong University of Petrochemical Technology

基　　金：广东省自然科学基金项目(2016A030307049);大学生创新创业训练与培育项目(201411656017;2015DCA004;2016py A033;2015py A041;2015py A042);大学生拔尖创新人才培养"培英计划"项目(广石化院[2015]21号)

摘　　要：以新浪微博为研究对象,基于MapReduce编程框架技术和PageRank评定方法,综合微博网络拓扑特征和用户行为因素,甄选出"直联热度"和"级联热度"等量化指标,用以表征评定微博用户影响力的动静态因素。构建了评估体系,提出一种基于混合量化指标与改进PageRank的微博用户影响力度量算法,并采用真实微博数据对该算法进行测试,对微博用户影响力进行了排名评估,与PageRank算法相比,表现出了较好的性能。Based on MapReduce＇s programming framework technology and the method for evaluation of PageRank,this article takes the sina microblog as the research object,synthetically considers the network topology characteristics and the factor of user behavior,selects the ＂direct heat＂and ＂cascaded heat＂as quantitative indicators to respectively represent dynamic and static factors in impact evaluation of microbloggers. On the basis of this,an assessment system is constructed,and then an algorithm is put forward to evaluate microbloggers＇ impact based on hybrid quantization index and improved PageRank algorithm. In addition,the experiment is carried out not only to test the algorithm performance but also to evaluate users＇ influence ranking on microblog. Compared with the PageRank algorithm,the result shows that the improved one has good performance.

关 键 词：新浪微博 用户影响力 混合量化指标 改进PageRank 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]