面向微博舆情演化分析的隐马尔科夫模型研究  

A Hidden Markov Model Research in the Microblog Public Opinion Evolutionary Analysis

作　　者：何建民[1,2] 李雪[1] 

机构地区：[1]合肥工业大学管理学院,安徽合肥230009 [2]过程优化与智能决策教育部重点实验室,安徽合肥230009

出　　处：《情报科学》2016年第4期7-12,共6页Information Science

基　　金：教育部人文社科研究规划基金项目(14YJA630015);国家自然科学基金项目(71271071)

摘　　要：微博舆情是互联网舆情表现的新形式,已成为影响国家安全和社会稳定的不确定因素,是网络舆情监测、预警和管控决策支持领域的新问题。本文运用隐马尔科夫模型(HMM)理论,选择网民特征、信息主题和信息内容完整度三维指标,根据微博舆情的事件演化阶段建立HMM状态空间、确定观测值、计算相关的参数,从而构建了面向微博舆情演化分析的隐马尔科夫模型,以揭示微博舆情演化发展的一般规律,并对模型的有效合理性实证验证。The microblog public opinion is a new expression form of internet public opinion, it has become the uncertainfactors that affect national security and social stability, and has become the new problem of network public opinion monitor-ing, early warning and control of decision-support areas. This paper use the theory of Hidden Markov Model, selectthree-dimensional indexes of netizens characteristics, information topics and content integrity. Base on the stages of theevolution of microblog public opinion envents to establish the HMM state space, determine observation, calculate the rele-vant parameters, thus build the HMM of microblog public opinion evolutionary analysis. It has reveal the general rule ofpublic opinion evolution, and empirical verify the effective and reasonable of the model.

关 键 词：网络舆情分析 隐马尔科夫模型 舆情分析模型 微博舆情监管 

分 类 号：G206.3[文化科学—传播学]