Educational data mining: Case study for predicting student dropout in higher education

Abstract

Reducing the student dropout rate in higher education is one of the challenges that universities are dealing with. By providing enriched study programs and qualified teams of professors, universities aim to enroll more students. Improving working conditions at the laboratories and other university resources aims to attract both ambitious students as well as high quality staff. After enrollment, the main goal of the faculty is to guide students into successful completion of their studies with the appropriate knowledge and skills acquired. Nowadays however, the development and deployment of Student Information Systems at the universities provides an appropriate infrastructure for student’s data organization and storage as well as data acquisition and deeper analyses. This data can help model the behavior of dropouts, and predict future dropouts, therefore giving chance to counselors to advise and guide students into success. This paper presents various data mining experiments and results obtained from data for the students from one of the faculties at the University Ss. Cyril and Methodius in Skopje. Initially, we give an overview of several data mining algorithms suitable for analysis of students’ data and dropout prediction. Furthermore, we explain modifications and applications of the algorithms over the existing student data. Finally, we provide a predictive model which will identify a subset of students who tend to drop out of the studies after the first year. The classification task aims to identify a pattern among students who tend to drop out.