基于主题挖掘和情感分析的“新冠肺炎疫情”舆情分析研究  

Public opinion analysis of COVID-19 epidemic using topic mining and emotion analysis

作　　者：杨秀璋 武帅 夏换[2] 于小民 Yang Xiuzhang;Wu Shuai;Xia Huan;Yu Xiaomin(School of Information of Guizhou University of Finance and Economics,Guiyang,Guizhou 550025,China;Guizhou Key Laboratory of Economics System Simulation of Guizhou University of Finance and Economics)

机构地区：[1]贵州财经大学信息学院,贵州贵阳550025 [2]贵州财经大学贵州省经济系统仿真重点实验室,贵州贵阳550025

出　　处：《计算机时代》2020年第8期31-36,共6页Computer Era

基　　金：贵州省科技计划项目“多源地理数据融合知识图谱构建方法在舆情分析中的应用——以贵州省为例”(黔科合基础[2019]1041);贵州省科技计划项目“基于大数据及图像识别的水族文献及濒危水书抢救性整理研究”(黔科合基础[2020]1Y279);贵州省科技计划项目“圆形地下连续墙结构时变性仿真研究”(黔科合基础[2019]1403);贵州财经大学2019年度校级项目“基于多源异构数据融合的建设项目管理知识图谱研究与实现”(2019XQN01)。

摘　　要：针对"新冠肺炎疫情"热点新闻和话题,提出一种基于主题挖掘和情感分析的舆情分析方法。通过Python抓取2020年1月20日至3月22日期间共计1389篇人民网的疫情新闻,利用数据预处理、特征词提取、词云可视化展现与"新冠肺炎疫情"相关的热点主题,再采用共词分析、LDA模型、知识图谱和情感分析算法挖掘舆情演化趋势。实验结果表明,此次肺炎疫情的情感呈现积极状态,热点主题包括疫情、防控、医院、工作、服务等。该方法能有效挖掘舆情事件的主题,具有一定的应用前景和使用价值。Aiming at the hot news and topics of COVID-19 epidemic,this paper proposes a public opinion analysis method using topic mining and emotion analysis.Python was used to capture a total of 1389 articles of epidemic news of People’s Daily Online from January 20 to March 22,2020.Data preprocessing,feature extraction and word cloud visualization were used to present the hot topics related to COVID-19 epidemic;and co-word analysis,LDA model,knowledge graph and emotion analysis algorithm were used to mine the evolution trend of public opinion.The experimental results show that the emotional state of COVID-19 epidemic is positive,and the hot topics include epidemic situation,prevention and control,hospital,work and service etc.,proving the method proposed in this paper can effectively explore the topic of public opinion events.

关 键 词：主题挖掘 情感分析 新冠肺炎疫情 知识图谱 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]