Regresar

Statistical learning to detect potential dropouts in higher education:Aa public university case study

Abstract:

Early detection of students tending to drop out is crucial to improve not only graduation rates but also education quality. By using basic statistical learning techniques, this work presents a simple way to pbkp_redict possible dropouts based on their demographic and academic characteristics. In order to reasonably pbkp_redict while gaining a better understanding of the dropout phenomenon, after a preliminary analysis, 4 classification methods are applied including 2 easy-Tointerpret ones. Some of the main results of this study show that almost 22% of current students are potential dropouts while being an older student and failing many subjects tend to cause dropout; on the other hand, passing more than 12 subjects and long-Term access to library materials can prevent students from leaving college.