Educational data mining to support identification and prevention of academic retention and dropout: a case study in introductory programming

Abstract

Several works in the literature emphasized data mining as efficient tools to identify factors related to retention and dropout in higher education. However, most of these works do not discuss if (or how) such factors may effectively contribute to decrease such rates. This article presents a data mining approach conceived to identify students at retention risk in a course of Intro to Computer Programming as well as guide preventive interventions to help such students to overcome this situation. Our results indicated an averaged predictive performance superior to 80% in both accuracy and F1 when identifying factors related to the retention. Moreover, during the two years of the project execution, the annual success rates in the course were the highest in comparison to the last five years.