A Data Science Approach to Predict the University Students at Risk of Semester Dropout

Abstract

In Bangladeshi institutions, the likelihood of student semester dropout has increased in recent years. A large number of university students, particularly in science background disciplines, are enrolled in a variety of undergraduate courses. Nevertheless, the perfection rate is poor. In general, students drop out for a variety of reasons, including academic, family, personal, and political concerns. The main focus of this study is to predict the risk of semester dropout in Bangladesh so that the massive dropout can be stopped. In this research, the current student information is preprocessed to discover the major reason as well as students whoever at the threat of semester dropout will help to grow a new structure in the area of higher education. To predict the dropout risk, random forest and logistic regression were practiced for obtaining the detection model.