电子学报 ›› 2017, Vol. 45 ›› Issue (11): 2800-2809.DOI: 10.3969/j.issn.0372-2112.2017.11.030

• 学术论文 • 上一篇    下一篇

基于多特征融合的微博用户权威度定量评价方法

张仰森, 郑佳, 唐安杰   

  1. 北京信息科技大学智能信息处理研究所, 北京 100101
  • 收稿日期:2016-11-06 修回日期:2016-12-16 出版日期:2017-11-25 发布日期:2017-11-25
  • 作者简介:张仰森,男,1962年6月出生于山西临猗,博士后,教授,研究方向为中文信息处理、人工智能.E-mail.zhangyangsen@163.com;郑佳,男,1991年10月出生于湖北松滋,硕士研究生,研究方向为中文信息处理、情感分析.E-mail.zhengjia0826@163.com;唐安杰,男,1990年11月出生于江苏盐城,硕士,研究方向为中文信息处理.E-mail:t_anjie@qq.com
  • 基金资助:
    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)

A Quantitative Evaluation Method of Micro-blog User Authority Based on Multi-Feature Fusion

ZHANG Yang-sen, ZHENG Jia, TANG An-jie   

  1. Institute of Intelligent Information Processing, Beijing Information Science and Technology University, Beijing 100101, China
  • Received:2016-11-06 Revised:2016-12-16 Online:2017-11-25 Published:2017-11-25

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.

关键词: 微博, 用户权威度, 用户影响力, UIRank, 层次分析法

Abstract: Micro-blog user authority is one of the important factors to evaluate the reliability of micro-blog information.In this paper,a quantitative evaluation model of micro-blog user authority is proposed based on multi-feature fusion.Firstly,the concept of user authority is proposed,which is defined by two parts.the user influence and the convinced degree.In the case that the user domain influence factor is not considered,we extracted four user characteristics which include the user information spread influence,the user information integrity degree,the user activity degree and the user platform authentication index based on the Sina Weibo data and construct the user authority quantitative calculation model.Then,we determine the weight of the four user characteristics based on the analytical hierarchy process and the extraction algorithms of them are given respectively.At the same time,we put forward the UIRank model based on the followed value and the follow relationship network between the users which is used to calculate the user information spread influence and proved to be more effective than the famous PageRank algorithm through experiments.The experimental results show that the method proposed in this paper is more reasonable to calculate the user authority of micro-blog user,and it provide a feasible scheme for the quantitative evaluation of the user authority.

Key words: micro-blog, user authority, user influence, UIRank, analytical hierarchy process

中图分类号: