Year: 2018
Type: conference-paper
Abstract: Massive open online courses (MOOCs) provide an opportunity for students to register for courses offered by best universities around the world.There are massive enrollments in the courses offered in MOOCs; however they suffer from low completion rate and low student retention.To address these problems, it is necessary to make early prediction of students' academic performance to enable targeted and timely interventions for those students at risk of non-completion.This study proposes the use of machine learning algorithms to predict students' academic performance in a MOOCs environment and compared the predictive power of four machine learning algorithms: Logistic Regression, Naïve Bayes, Random Forest and K Nearest Neighbor.Our results show that performance of predictive models are promising and can be used for early prediction of students that are likely to fail in a course.Furthermore, Random forest classifier outperforms other classifiers in a statistical significant way.
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: gold
OpenAlex ID: W4252659442