Predicting the likelihood of dropping out from MOOC-based program

 

This research aims to identify the likelihood of dropping out from a MOOC-based program. Program dropout happens both, within and between courses. We identify the key dropout factors and develop a machine learning model to predict future student dropout.

Contribution:

Results from our prediction models reveal that learners are more likely to leave a MOOC-based program after the first few courses (1-3). Grades, level of education and gender are highly correlated with learner dropout. We suggested some potential initial interventions which could be highly valuable as a starting point.

Research Publications

Research Team

Dr. Chris Caplice

Dr. Chris Caplice

Senior Research Scientist

Dr. Eva Ponce

Dr. Eva Ponce

Director

Andrea Moreno

Andrea Moreno

Research Assistant