基于多特征融合的微博用户权威度定量评价方法

发布时间：2018-09-14 15:33

【摘要】：微博用户权威度是评价微博信息可靠性的重要因素之一.本文针对微博用户权威度的定量计算提出了一种基于多特征融合的微博用户权威度定量评价模型.首先,提出了用户权威度的概念,将其定义为用户影响力和被信服度两部分组成;在暂不考虑用户领域影响因子的情况下,基于新浪微博数据,抽取出微博用户信息传播影响力、用户信息完整度、用户活跃度以及用户平台认证指数4项评价特征,以构建了用户权威度定量计算模型;然后,采用层次分析法对所构建模型的4项评价特征的权值进行确定,并分别给出了4项评价特征的提取算法.同时,在用户关注关系网络的基础上,提出了一种基于用户被关注价值的用户信息传播影响力模型UIRank,并通过实验验证了其比PageRank算法更加有效.实验结果表明,本文提出的微博用户权威度定量计算模型比较合理,为用户权威度的定量评价提供了一种可行的解决方案.
[Abstract]:Weibo user authority is one of the important factors to evaluate the information reliability of Weibo. This paper presents a quantitative evaluation model for Weibo user authority based on multi-feature fusion. First of all, the concept of user authority is proposed, which is defined as user influence and credibility. In order to construct a quantitative computing model of user authority, four evaluation features of Weibo's influence on information dissemination, user information integrity, user activity and user platform authentication index are extracted. The weights of the four evaluation features of the constructed model are determined by using the analytic hierarchy process (AHP), and the extraction algorithms of the four evaluation features are given respectively. At the same time, on the basis of user concern relation network, a user information dissemination influence model, UIRank, is proposed based on user concerned value, and it is proved to be more effective than PageRank algorithm through experiments. The experimental results show that the quantitative calculation model of Weibo's user authority is reasonable, which provides a feasible solution for the quantitative evaluation of user authority.
【作者单位】：北京信息科技大学智能信息处理研究所;
【基金】：国家自然科学基金(No.61070119,No.61370139,No.61772081,No.61602044) 北京市属高等学校创新团队建设与教师职业发展计划(No.IDHT20130519)
【分类号】：TP393.092

【相似文献】