李彤彤1 郭栩宁1 周彦丽1 李 坦2
（1.天津师范大学 教育学部，天津 300387；2.北京师范大学 教育学部，北京 100875）
【摘 要】新冠肺炎疫情对身心尚处于发展阶段的大学生群体造成了一些消极影响，亟须引起关注。随着社
交媒体的广泛应用以及大数据技术的发展，微博已经成为大学生记录生活、表达观点、分享交流等活动的主要途
径，通过微博文本挖掘可以实时地了解大学生群体真实的观点和情感状态。文章立足于深度挖掘微博文本，首
先，结合疫情的演变情况和高校系统的应对方式，将疫情时期划分为疫情暴发期、停课不停学初期、停课不停学
适应期、复学初期、复学后常态化防控期五个阶段，采用支持向量机算法对微博文本进行情感分析，将其逐条划
分为“积极”“消极”两类情感倾向。其次，通过构建LDA主题模型以及结合TF-IDF特征挖掘，分析不同时间阶段
消极情感倾向微博文本中隐含的主题。研究发现，学生整体情感状态偏向于积极，但会随疫情与教学的变化而产
生波动，其消极情感的产生与疫情演变、校园管理、学业压力以及就业压力密切相关。通过探究疫情期间某高校
大学生情感状态及变化趋势并探寻引发学生消极情感的相关事件，以期为高校教育工作者有效引导、干预、管理
学生提供依据和参考。
【关键词】新冠肺炎疫情；大学生；情感状态；文本情感分析；文本挖掘；文本主题挖掘
College Students’ Emotion State and Its Influencing Factors under COVID-19: Evidence from
Micro-blog Mining
LI Tongtong1, GUO Xuning1, ZHOU Yanli1
and LI Tan2
(1. Department of Education, Tianjin Normal University, Tianjin 300387, China; 2. Faculty of Education, Beijing Normal University,
Beijing 100875, China)
Abstract: The COVID-19 has caused negative effects on college students who are still in the stage of physical and mental development,
which deserves urgent attention. With wide application of social media and the development of big data, micro-blog has become college
students’ routine to record daily lives, to express opinions, to share and to communicate through micro-blog. Micro-blog text mining
shows its advantage in understanding the real opinions and emotion states of college students in real time. Firstly, the COVID-19 period
was divided into five stages according to the evolution of epidemic situation and how university system responded: the outbreak period,
the initial period of suspension and non-stop schooling, the adaptation period of suspension and non-stop schooling, the initial period of
resumption of classes, and the normal prevention and control period after resumption of classes. Support Vector Machine (SVM)
algorithm was used for text sentiment analysis. Micro-blog text is divided into “positive”and“negative” emotion tendency item
by item. Secondly, LDA theme model is constructed and combined with TF-IDF feature mining. Hidden themes in micro-blog text with
negative emotion tendency in different stages are analyzed. The main findings are: overall emotion states of students tended to be positive, but
fluctuated with the changes of the epidemic periods and teaching mode, and students’negative emotions were closely related to evolution of
the epidemic, campus management, academic pressure and employment pressure. The emotion states and change trends of college students
during the epidemic period are explored, as well as related events causing students’negative emotions, in order to provide basis and reference
for educators in effectively guiding, intervening and managing college students.
Keywords: COVID-19; college students; emotion state; text sentiment analysis; text mining; text theme mining
下载： 
新冠肺炎疫情下大学生情感状态及其影响因素分析.pdf
