Computer Science Students Early Drop-Out Prediction Using Machine Learning: A Case Study

Abstract

AbstractHigh dropout rates in computer science programs are a major global issue. In this chapter, we describe the results of the educational machine learning case study with the aim of early prediction of computer science students’ dropout in the Virumaa College of Tallinn University of Technology. For many years TalTech Virumaa College has faced a high dropout rate among first-year students. Among computer science students it is about 40%. In order to reduce the dropout rate, interviews and surveys with student candidates are carried out in the College admission process since the academic year 2019/20 to assess students’ motivation and readiness to study and ensure the suitability of the specialty. To determine factors that influence attrition rates we analyze the data collected during the admission process and apply the text mining techniques to analyze the students’ essays conducted at the end of the first-year study in the frame of the course Introduction to the Specialty. Using these factors, we evaluate predictive models and therefore make recommendations to improve the admission process and reduce the drop-out rate for first-year students.