基于MOOC课程评论的话题挖掘与情感分析研究

Topic Mining and Sentiment Analysis Based on MOOC Course Reviews

为了深入挖掘与分析在线课程评论文本,探索学习者参与在线课程学习时关注的话题及其情感态度,为提高在线课程质量提供帮助。首先采用词频分析方法,实现对学习者在线课程评论内容的整体认识;然后利用非监督学习方法潜在狄利克雷分布主题模型对评论文本信息的特征结构、语义内容进行自动挖掘和分析,得到学习者的关注话题;最后对每个话题的课程评论文本进行情感倾向分析,得到学习者的情感倾向分布。实验结果表明,在参与课程学习的过程中,学习者主要关注教师授课、课程内容和学习资源3个话题。情感分析结果显示,学习者对于该课程普遍表示满意和赞赏,但是对于该课程学习资源表达了较多负面情感。

In this paper,the review text of online courses is deeply mined and analyzed with the aim to explore the topics and emotional attitudes that learners are interested in when participating in online courses,and help improve the quality of online courses.Firstly,the word frequency analysis method is adopted to overall understand the review content of learners’online courses,and then the LDA topic model is used to automatically mine and analyze the characteristic structure and semantic content of the review text information,so as to obtain the topic of learners’attention.Finally,the text of course review on each topic is analyzed to get the distribution of learn⁃ers’sentiment tendency.The experimental results show that in the process of participating in the course learning,learners mainly pay attention to the teachers’teaching,course content and learning resources.The result of sentiment analysis shows that learners general⁃ly express satisfaction and appreciation for the course,but express negative feelings on the learning resources of the course.