A TWO-LAYER CASCADING METHOD FOR DROPOUT PREDICTION IN MOOC

Bowei Hong, Zhiqiang Wei, and Yongquan Yang

Keywords

MOOC, machine learning, dropout prediction, ensemble of classifiers

Abstract

Distance education has surged tremendously with the gradual progress in information technology over the past decade. There has been dramatically increasing online learning platforms showing up in daily life. However, while benefiting from these online learning platforms, people also come across issues involved with them, such as high dropout rate. To solve these issues, various works have been done, including the analysis and prediction of dropout. This paper studies dropout prediction for Massive Open Online Course (MOOC) using learning activity information of learners. A two-layer cascading classifier is applied for prediction, which is a combination of three different machine learning classifiers (i.e., Random Forest, Support Vector Machine, and MultiNomial Logistic Regression). Experimental results indicate that the technique is promising in predicting dropouts with achieving 97% precision.

Important Links:



Go Back