Predicting dropout-prone students in e-learning education system

Abstract

High rate of students dropout in courses has been a major problem for many universities or educational institutions that offer online education. If the dropout-prone students can be identified in their early stages, the dropout rate can be reduced by providing individualized care to the students at-risk. Due to the electronic nature of the e-learning courses, various attributes of the student progress can be monitored and analyzed automatically over time. In this paper, a technique for predicting students who are prone to dropout from the online courses has been proposed that progressively analyzes a set of per-learner attributes of the students' activities overtime. Since a single machine learning technique may fail to accurately identify some dropout-prone students whereas others may succeed, this technique uses a combination of multiple classifiers (ensemble of classifiers) for this analysis. The results of the validation found the technique to be promising in predicting dropout-prone students.