作　　者： (刘发升）; (韩青菊）;

机构地区： 江西理工大学信息工程学院,赣州341000

出　　处： 《科学技术与工程》 2017年第22期243-248,共6页

摘　　要： 针对当前微博影响力度量算法中多集中于用户行为属性,忽略博文、结点本身价值的问题,从微博用户信息出发,以线性加权模型为基础,综合分析用户的行为属性、博文相似度、节点相似度,创建影响力评价指标体系。利用Page Rank算法思想,提出了基于用户行为和博文内容的用户影响度量模型(user influence measurement rank,UMR)。通过采用新浪微博真实数据集测试,计算用户的影响力,验证了UMR算法在博文内容的基础上,能客观地反映用户的交互行为,消除僵尸用户对排序的影响,因而更科学、更合理。 In terms of algorithms of the influence measurement in which the users＇ behavior attributes are more concentrated on as well as the value of blog content and the node itself are ignored in the current micro-blog,the author,from the perspective of the microblog user information,make a comprehensive analysis of the users＇ behavior attributes and of content and node similarity to create an influence evaluation index system based on the linear weighting model.By means of Page Rank algorithm,the author propose User Influence Measurement Rank—UMR on the basis of users＇ behaviors and contents.By adopting the authentic data of Xinlang micro-blog to test and calculate the user＇s degree of influence,the author make a verification of the UMR algorithm based on blog contents that it can objectively reflect the users＇ interactions and can eliminate the zombie users＇ influence upon sorting,thus becoming more scientific and more reasonable.

关 键 词： 微博 线性加权 算法 用户影响力 博文内容