Evaluating On-Line Courses via Reviews Mining

摘要: Because the participants are not limited by age-, gender-, race-, or geography-related barriers, recently, massive open online courses (MOOC) have witnessed remarkable growth in number of self-learners, providers and platforms. MOOC learners usually share some learning experiences release millions course-related comments discussion forum. On one hand, these could reflect learners’ attitudes toward courses. other semantic knowledge hidden would assist to choose appropriate help instructors improve their courses’ attraction. Recently, few research works focus on evaluating through reviews mining. Thus, this paper constructs a curriculum evaluation system based reviews, which quantifies from different topics. Firstly, we employ latent dirichlet allocation (LDA) mine generated students, obtain topic-word distribution matrix comment-topic can describe topics course comments. Next, emotion values each topic calculated auto-encoder Bi-LSTM text classification model. We utilize emotions quantified scores establish comprehensive system. The experimental results show that there five main indicators abstracted students’ instructor, content, assessment, platform, hot Moreover, comment texts under objectively accurately converted into numerical marks, provide students educators with reliable references.