A Survey of Student Dropout Analysis Using Machine Learning Approach

Abstract

Online learning and conventional learning are two important methods that students pursue their educational degrees or expand their knowledge. The enormous rate of dropout of online students and the fact that this rate is still increasing leads to the concern of raise student retention rate. And the large number of conventional student dropout create loss in economics, time and education resources. Therefore, it is important to seek the factor affecting student dropout and investigate effective machine learning-based models on prediction of student dropout. Different factors including the clickstream, academic information, family are discussed, followed by introductory information about current popular machine learning algorithms. According to this comprehensive review, the grade of online learners and the total number of students assessment to courses seem to be the most powerful features and conventional students are affected by social contact and attendance of social activities.