基于PageRank的微博用户影响力评估模型研究

摘要

为解决PageRank算法对社交网络中的用户进行影响力评估时产生的“主题偏移”和PR值均匀分配问题,文中以新浪微博为例,构建了用户活跃度、人气值、兴趣度和用户价值四个评估指标,将用户的自身属性作为影响力的评定因子和分配因子,从内在和外在两方面综合考虑用户的影响力,基于PageRank算法提出一种对用户影响力进行综合评估的CIRank模型.在真实微博数据集下,将CIRank模型与传统影响力评估模型结合新浪官方影响力排名进行比较分析,验证其准确性和可行性.%In order to solve the problem of “theme deviation” and the average distribution of PR value when the PageRank algorithm is used to evaluate the influence of the users in the social network,this paper constructs four evaluation indicators as users' activity,popularity,interest and user value based on Sina microblog,the users' own attribute is used as the evaluation factor and distribution factor of influence and it considers the influence of users from the external and internal aspects,the CIRank model is proposed to evaluate the user influence comprehensively based on the PageRank algorithm,and the CIRank model is compared with the traditional influence evaluation model and Sina's official ranking to verify its accuracy and feasibility in the real microblog data.