基于LDA模型的高校师德舆情演化及路径传导研究

基于LDA模型的高校师德舆情演化及路径传导研究

长春工业大学

Research on the Evolution and Path of Public Opinions about Teachers' Ethics in Colleges and Universities Based on LDA Model

摘要/Abstract

摘要：

【目的/意义】构建高校师德舆情微博用户评论LDA模型，对舆情演化特征的识别及该话题下的网络主题传播路径分析，可以更精准识别舆情演化特征和分析关键主题传播路径，帮助高校和相关部门更为有效地进行舆情监管和舆情引导。【方法/过程】本文以“天津大学一教授学术造假”事件为例，基于LDA模型构建高校师德舆情下微博用户主题生成模型，采用困惑度评价指标确定LDA模型最优主题数，采用信息熵确定每一主题在不同日期的主题强度，通过关键词共现知识图谱、词云展现舆情话题的演变，最后基于主题相似度确定主题传播路径。【结果/结论】基于LDA模型的微博用户评论主题生成模型和共现知识图谱可以解析出网络用户群体关注的重要主题热点，LDA模型和信息熵能够精准识别舆情演化特征，识别主题最优传播路径进行舆论引导。【创新/局限】文章创新性地构建高校学术道德舆情的LDA主题模型，有效确定微博用户群体主题、识别舆情演化特征、分析主题间传播路径，具有普适性，进一步扩大高校师德其他舆情分析及结合网络舆情情感分析为下一步的研究内容。

关键词:

LDA, 微博, 高校师德, 舆情演化, 路径传导

Abstract:

【Purpose/significance】Constructing the LDA model of user reviews of university teachers’ morals and public opinion on Weibo, the identification of the evolution characteristics of public opinion and the analysis of the network topic propagation path under this topic, can more accurately identify the evolution characteristics of public opinion and analyze the propagation path of key topics, helping universities and education The department conducts public opinion supervision and public opinion guidance more effectively.【Method/process】This article takes the event of "a professor of Tianjin University academic fraud" as an example. Based on the LDA model, the topic generation model of Weibo users under the public opinion of teachers’ ethics in colleges and universities is constructed. The confusion evaluation index is used to determine the optimal number of topics in the LDA model, and information entropy is used to determine each The topic intensity of the topic on different dates is displayed through the keyword co-occurrence knowledge graph and word cloud to show the public opinion topic, and finally the topic transmission path is determined based on the topic similarity.【Result/conclusion】The microblog user comment topic generation model and co-occurrence knowledge graph based on the LDA model can analyze the important topic hotspots of the network user community. The LDA model and information entropy can accurately identify the evolutionary characteristics of public opinion, identify the optimal propagation path of the topic, and guide the public opinion.【Innovation/limitation】The article innovatively constructs the LDA topic model of academic moral public opinion in colleges and universities, effectively determines the subject of Weibo user groups, identifies the evolutionary characteristics of public opinion, and analyzes the communication path between topics. It is universal, it further expands the analysis of other public opinions of college teachers’ ethics and the integrated network Public sentiment analysis is the next research content.

Key words:

LDA, Weibo, College Teachers' Ethics, Public opinion evolution, Path conduction

张雷谭慧雯张璇韩龙.

基于LDA模型的高校师德舆情演化及路径传导研究 [J]. 情报科学.