计算机工程 ›› 2019, Vol. 45 ›› Issue (12): 294-299.doi: 10.19678/j.issn.1000-3428.0053029

• 开发研究与工程应用 • 上一篇    下一篇

一种改进的微博用户影响力评估算法

黄贤英, 阳安志, 刘小洋, 刘广峰   

  1. 重庆理工大学 计算机科学与工程学院, 重庆 400054
  • 收稿日期:2018-10-30 修回日期:2018-12-21 发布日期:2018-12-28
  • 作者简介:黄贤英(1967-),女,教授,主研方向为社交网络、传播模型;阳安志(通信作者),硕士研究生;刘小洋,副教授、博士后;刘广峰,硕士研究生。
  • 基金项目:
    国家社会科学基金(17XXW004);教育部人文社科青年基金(16YJC860010);重庆市教育委员会人文社会科学研究项目(17SKG144,18SKGH100);2018年重庆市科委技术创新与应用示范项目(cstc2018jscx-msybX0049);2017年度重庆市高校网络舆情与思想动态研究咨政中心开放课题(KFJJ2017024)。

An Improved Algorithm for Microblog User Influence Evaluation

HUANG Xianying, YANG Anzhi, LIU Xiaoyang, LIU Guangfeng   

  1. College of Computer Science and Engineering, Chongqing University of Technology, Chongqing 400054, China
  • Received:2018-10-30 Revised:2018-12-21 Published:2018-12-28

摘要: 在已有PageRank算法构建的微博用户影响力评估模型中,存在用户自身属性信息欠缺以及在用户不活跃期间其影响力被误判下降的问题。为此,综合考虑用户自身的属性,基于用户的活跃度、认证信息及博文质量来确定其自身的基本影响力,通过引入用户博文的传播率挖掘用户的潜在影响力,结合用户不同好友的质量,基于改进的PageRank算法构建微博用户影响力评估算法。实验结果表明,与改进BWPR算法相比,该算法准确率、召回率和F值分别提高13.5%、10.1%和12.3%,能准确、客观地反映微博用户的实际影响力,可为社交网络中的意见领袖挖掘、信息传播和舆论引导等研究提供参考。

关键词: 微博, 社交网络, 影响力, PageRank算法, 传播率

Abstract: The current PageRank algorithm based microblog user influence evaluation model has many problems,such as the lack of users' attributes and the incorrect evaluation of users' influence during their inactive period.In this paper,the attributes of users are comprehensively considered,and their basic influence is confirmed based on users' activeness,authentication information and blog quality.Then,according to the blog dissemination rate,this paper further excavates the potential influence of users.Ultimately,with the evaluation of the quality of different microblog friends,a microblog user influence evaluation algorithm is established on the basis of improved PageRank algorithm.Experimental results show that compared with BWPR improved algorithm,the accuracy,recall rate and F value of the proposed method are increased by 13.5%,10.1% and 12.3% respectively.It can reflect the actual influence of users in a more accurate and objective way,which can provide references to the researches on opinion leader identification,information dissemination and the guidance of public opinions in social networks.

Key words: microblog, social network, influence, PageRank algorithm, dissemination rate

中图分类号: