面向MOOC课程评论的学习者话题挖掘研究  

Study on Learners' Topics Mining of MOOC-oriented Course Review

作　　者：刘三女牙[1] 彭晛 刘智[1] 孙建文[1] 刘海[1] 

机构地区：[1]华中师范大学国家数字化学习工程技术研究中心,湖北武汉430079

出　　处：《电化教育研究》2017年第10期30-36,共7页E-education Research

基　　金：2016年教育部-中国移动科研基金项目"国家教育大数据相关问题研究"(项目编号:MCM20160401);2016年度教育部人文社会科学研究青年基金项目"高校慕课环境下的互动话语行为及其对学习效果的影响机理研究"(项目编号:16YJC880052)

摘　　要：研究以果壳网MOOC学院的"财务分析与决策"课程为实验对象,通过分析课程评论帖进行学习者话题的挖掘。文章不仅采用了高频词汇分析的定量方法,实现对学习者课程评论内容的整体认识,并且,根据参与评论学习者的课程完成情况,分别对已完成和未完成两种类型的学习者展开定性的学习分析研究,应用非监督学习方法 LDA模型自动挖掘和解析文本评论信息的特征结构和语义内容,并探究和追踪学习者关注的热点话题演化趋势。实验结果表明,学习者认可和赞赏了该门课程,并且尤为关注课程内容以及教师授课形式话题;相比课程完成者,未完成者更倾向于解释其未完成课程的主要原因,表达出更为消极的话题内容,并较少涉及课程本身相关的专业理论知识。This study takes the course Financial Analysis and Decision-making of MOOC college in Guokr as the experimental subject to mine the learners＇ topics through analysis of course review posts.Firstly, the study adopts the quantitative method of high frequency words analysis to realize the overall understanding of the content of learners＇ course review. Then, the learning analysis is used to study the learners who have completed the course as well as learners who haven＇t completed it respectively. The unsupervised learning method LDA model is employed to automatically excavate and resolve the feature structure and semantic content of text review information, explores and tracks the trend of hot topics that learners are concerned about. The study results show that learners highly recognize and appreciate this course, and pay special attention to the course content as well as teachers＇ teaching forms. Compared to the completer, the learners who haven＇t completed the course tend to explain the main reasons for the unfinished course, express more negative topics and less professional theoretical knowledge of the course.

关 键 词：MOOC 文本评论 话题挖掘 LDA 

分 类 号：G434[文化科学—教育学]