Assisted education: Using predictive model to avoid school dropout in e-learning systems

Abstract Dropping out of school is a real challenge for educational specialists. Distance education classes deal with a huge number of students’ disengagement with social and economic costs. Behavioral, cognitive, and demographic factors may be associated with early school dropout. This paper proposes an enhanced machine learning ensemble predictive architecture, capable of predicting the disengagement of students along with the class. It notifies teachers, enabling them to intervene effectively and make students’ success possible, and students to give them a chance to turn back. To evaluate the proposal, a case study showed the feasibility of the solution and the use of its technologies. Results pointed out a significant increase of gain in accuracy along with the class, reaching 93% of precision at the end.