突发事件网络舆情趋势的影响因素分析及预测  

Analysis and Prediction of Factors InfluencingOnline Public Opinion Trends in Sudden Incidents

作　　者：徐孟圆 张旖华 XU Meng-yuan;ZHANG Yi-hua(School of Physics and Information Technology,Shaanxi Normal University,Xi'an710119,Shaanxi;School of Language and Literature,Shaanxi Normal University,Xi'an710119,Shaanxi)

机构地区：[1]陕西师范大学物理学与信息技术学院,陕西西安710119 [2]陕西师范大学文学院,陕西西安710119

出　　处：《商洛学院学报》2024年第6期51-57,共7页Journal of Shangluo University

摘　　要：以微博平台为研究对象,聚焦突发事件的网络舆情传播,旨在识别和量化影响舆情趋势的关键因素,并利用机器学习技术构建精准预测模型。通过网络爬虫技术收集27 422条相关微博数据,提取内容特征、用户互动行为等多维特征,采用聚类算法分类微博内容,结合随机森林算法预测转发量。结果表明,评论数和点赞数对舆情传播的影响最显著,模型预测准确率高达87%,为优化舆情应对策略提供科学依据。This study focuses on the Weibo platform to investigate the dissemination of online public opinion during sudden incidents.It aims to identify and quantify key factors influencing opinion trends and develop an accurate prediction model using machine learning techniques.A total of 27422 related Weibo posts were collected through web crawling,and multidimensional features such as content characteristics and user interaction behaviors were extracted.Clustering algorithms were applied to classify the content of the posts,and the random forest algorithm was used to predict the number of reposts.The results indicate that comment count and like count are the most significant factors influencing public opinion dissemination.The prediction model achieved an accuracy of 87%,providing a scientific basis for optimizing public opinion response strategies.

关 键 词：爬虫 机器学习 聚类算法 随机森林 

分 类 号：TP393.092[自动化与计算机技术—计算机应用技术] C912.63[自动化与计算机技术—计算机科学与技术]