How to Tell Successful and Unsuccessful Online Learners: A "Grey-Box" Approach to Student Profiling

• In Event: Machine Learning Applications in Education

Abstract

Student profiling based on Learning Management System tracking behaviors is an emerging topic and has had a strong impact on promoting students’ learning experiences in massive online courses. The past studies have contributed to profiling online behaviors; however, little has been done to develop a systematic student profiling solution that balances education theories with computation models. Besides, there is limited research investigating what behavior patterns lead to successful performance and how learners behave differently between successful and unsuccessful learning groups. To address the gaps in research, we analyzed the behaviors of 26,862 online learners from an open university by applying learning analytical techniques. Specifically, we developed an interpretable student profiling solution and identified behavior patterns between higher and lower performers.

Authors

• Hui Shi, Florida State University
  Presenting Author

• Yihang Zhou, Tongji University
  Non-Presenting Author