开放学习研究

在线课程与教学

当前位置: 首页 >> 期刊栏目 >> 在线课程与教学 >> 正文

面向MOOC课程评论的主题挖掘与情感分析研究

2021年第4期  点击:[]


王洪鑫1 闫志明2 陈效玉1 张铭锐2

(1.鲁东大学 教师教育学院,山东 烟台 264025;2.鲁东大学 教育科学学院,山东 烟台 264025)

【摘 要】MOOC为各类学习者的能力提升提供了良好的平台,成为目前较为流行的学习方式。学习者完成课程学习,将学习体验以文本的形式反馈给教育者,产生了大量的文本数据。针对教育者对该类数据查看不及时、不全面从而导致课程缺陷不能及时弥补的问题,构建了一种面向MOOC课程评论的主题挖掘与情感分析模型。该模型利用词云对课程评论进行整体分析,运用LDA模型分析文本内容的语义信息和特征结构,自动挖掘学习者发表的课程评论中的隐含主题,最后借助字符级CNN算法训练深度情感分析模型,分析文本信息的情感极性,得到不同主题下学习者的情感分布。研究将该模型应用到“面向核心素养的信息化教学设计”课程中,结果发现:该模型能较好地挖掘出学习者在课程学习时的关注主题(如授课方式、评价方式等),并分析相应主题的情感极性,帮助教育者总结出MOOC课程评论中学习者对课程本身的意见与建议。最后,针对分析结果,给出MOOC建设的相关建议。

【关键词】MOOC;LDA;主题挖掘;情感分析

Research on Topic Mining and Emotion Analysis for MOOCs Course Review

WANG Hongxin1, YAN Zhiming2, CHEN Xiaoyu1 and ZHANG Mingrui2

(1. School of Teacher Education, Ludong University, Yantai 264025, China; 2. School of Education and Science, Ludong University, Yantai 264025,China)

Abstract:Providing a platform for all kinds of learners to improve their ability, MOOCs have become a popular way of learning. Learners complete the course, and give feedback to the educators in forms of text, producing a large amount of text data. Bearing the issue of delayed feedback may lead to delayed course defects remediation in mind, a topic mining and emotion analysis model for MOOC course commentary is proposed. This model makes use of word cloud to analyze the whole course review, uses LDA model to analyze the semantic information and feature structure of the text content, and automatically mine the hidden topics in the course review published by learners. Finally, the character-level CNN algorithm is used to train the deep affective analysis model, and the affective polarity of text information is analyzed to get the affective distribution of learners under different topics. This model is applied to “information-based instructional design for core literacy” course, and the results show that this model could dig out learners’ attention topics (such as teaching methods, evaluation methods, etc.), as well as analyzes the emotional polarity of the corresponding topics, and assists educators to summarize the learners’ opinions and suggestions on the MOOC course review. According to the analysis results, this article gives some suggestions on MOOCs construction.

Keywords:MOOC; LDA; topic mining; affective analysis

下载:  面向MOOC课程评论的主题挖掘与情感分析研究.pdf


关闭

最新文章New
我要投稿
热点文章Hot