电信科学 ›› 2017, Vol. 33 ›› Issue (9): 28-35.doi: 10.11959/j.issn.1000-0801.2017204

• 研究与开发 • 上一篇    下一篇

基于干扰相似度的多话题演化模型

陈叶斐,张学军,黄卫东   

  1. 南京邮电大学,江苏 南京 210023
  • 修回日期:2017-06-28 出版日期:2017-09-01 发布日期:2017-09-11
  • 作者简介:陈叶斐(1973-),女,南京邮电大学助理研究员,主要研究方向为网络舆情预警。|张学军(1969-),男,博士,南京邮电大学教授,主要研究方向为智能信息处理、复杂网络与系统和认知网络频谱感知等。|黄卫东(1968-),男,博士,南京邮电大学教授,主要研究方向为应急管理、数字化预案和网络舆情分析。
  • 基金资助:
    国家自然科学基金资助项目(71671093);江苏省高校哲学社会科学基金资助项目(2015SJB018);南京邮电大学人文社会科学基金资助项目(NYS214028)

Multiple topics evolution model based on similarity of interference

Yefei CHEN,Xuejun ZHANG,Weidong HUANG   

  1. Nanjing University of Posts and Telecommunications,Nanjing 210023,China
  • Revised:2017-06-28 Online:2017-09-01 Published:2017-09-11
  • Supported by:
    The National Natural Science Foundation of China(71671093);Philosophy and Social Science Foundation of Education Department of Jiangsu Province(2015SJB018);Humanities and Social Science Foundation of NUPT(NYS214028)

摘要:

当前演化模型研究中,主要是单一话题在网络中的传播,较少考虑多话题之间的相互影响因素。在SIR 模型的基础上提出了基于干扰相似度的多话题演化模型,该模型中的干扰是通过话题相似度对传播概率的影响来表征的。仿真结果表明,在临界值以内,正负两种趋向的话题相似度分别对话题演化的进程起到加强或阻碍作用,作用程度随着被干扰节点的度而变化,分别表现为正向相似度下的演化一致性和负向相似度下的演化分离性。超过临界值时,加强或阻碍作用均趋于饱和。

关键词: 话题演化, 多话题, SIR模型, 干扰相似度

Abstract:

The current researches of evolution model mainly focus on the spread of the individual topics,rarely considering the influential factors between multiple topics.A new topic evolution model was proposed by considering the interference among topics based on SIR model,which characterized by the influence of the similarity of the topic on the probability of propagation.The experimental results show that within the critical value,the similarity degree of positive and negative trends enhance or hinder the process of topic evolution,and the degree of action varies with the degree of interference nodes,which is expressed as evolutionary consistency under positive similarity and the evolutionary separability under negative similarity.When the critical value is exceeded,the effect of strengthening or hindering tends to saturation.

Key words: topic evolution, multiple topics, susceptible-infective-removal model, similarity of interference

中图分类号: 

No Suggested Reading articles found!