Negative Sentiment Analysis of MOOC Comments Based on Machine Learning

Shuhan Liu,Genfu Yang
DOI: https://doi.org/10.1007/978-981-15-1864-5_77
2020-01-01
Abstract:As the MOOC learning scenarios become more diverse, users also generate more unstructured text user data during use. This study takes the five computer courses of the Chinese University MOOC network as the experimental object, and mines a total of 20,176 comments, combined with the machine learning model and high-frequency lexical analysis method to analyze the negative comments. The study found that negative comments talked about the following five topics, course content, course difficulty, teacher attributes, supporting resources, and interaction design. And each topic has a more concentrated opinion. For example, the speed of the teacher, the amount of content, the course is difficult. Based on the above research results, both the course provider and the MOOC platform can be optimized to achieve better user evaluation and development.
What problem does this paper attempt to address?