基于偏差规则马尔可夫模型的网络舆情情感预测研究  

Emotion Prediction of Network Public Opinions Based on the Deviation Rules Markov Model

作　　者：史伟 薛广聪 何绍义 Shi Wei;Xue Guangcong;He Shaoyi(School of Economics and Management,Zhejiang Ocean University,Zhoushan 316022;School of Information Engineering,Huzhou University,Huzhou 313000;College of Business and Public Administration,California State University,San Bernardino,San Bernardino 91708)

机构地区：[1]浙江海洋大学经济与管理学院,舟山316022 [2]湖州师范学院信息工程学院,湖州313000 [3]加州州立大学圣伯纳迪诺分校商业与公共管理学院,圣伯纳迪诺91708

出　　处：《情报学报》2023年第9期1065-1077,共13页Journal of the China Society for Scientific and Technical Information

基　　金：国家社会科学基金一般项目“重大突发事件中网民情感状态演变规律及引导研究”(20BXW013)。

摘　　要：针对微博短文本评论,基于情感分析技术,从情感类转移等角度实现对网络舆情演变趋势的预测。本文以突发事件“新冠肺炎疫情”初期的相关微博评论文本作为研究对象,基于扩展关联规则Apriori算法和马尔可夫链提出一种新的方法,即偏差规则马尔可夫模型(the deviation rules Markov model,DRMM)。该模型分析了网民情感类间的相关性和转移性,通过计算不同的情感类转移概率、构建时变的情感状态转移矩阵对疫情初期网民情感状态的变化趋势进行预测。实验采取平均绝对误差(mean absolute error,MAE)和均方根误差(root mean squared error,RMSE)来衡量模型预测值与真实值之间的误差。研究结果表明,该模型具有较好的有效性和准确性,预测值和真实值的拟合效果在预期范围之内。Using sentiment analysis technology,the evolution of online public opinion can be predicted from the perspective of emotion transfer in short microblog comments.In this study,the short texts of related microblog comments in the early stage of the COVID-19 pandemic were taken as the research object.Based on the extended association rule Apriori Algorithm and Markov Chain,a new method called the deviation rule Markov model is proposed.This model analyzes the correlation and transfer between Internet users’emotion classes and predicts the changing trends of Internet users’emotional states in the early stage of the pandemic by calculating the transfer probability of different emotion classes and constructing a time-varying emotion state transfer matrix.The experimental results demonstrated that the emotional state of netizens after the pandemic outbreak was not negative but gradually changed to“positive”emotions over time.Through a comparative analysis of examples,the validity and accuracy of the affective prediction model for online public opinion proposed in this study were verified.

关 键 词：情感转移 网络舆情 短文本挖掘 关联规则 马尔可夫链 

分 类 号：C912.63[经济管理] G206[社会学]