Assessment of a predictive model for academic performance in a small-sized programming course

Abstract

Programming courses often demonstrate poor academic performance. To address this challenge, one strategy involves developing predictive models to identify students facing difficulties and implementing early interventions during the course. However, additional research is necessary to evaluate the effectiveness of these predictions and interventions. The aim of this study is to evaluate the quality of an academic performance prediction model in a programming course with few students, which proves to be a complex process due to the limited number of records. The methodology employed both a predictive model and a questionnaire focusing on motivating learning strategies. Results show that the proposed model achieves 86% accuracy in its predictions, beginning from week 7 of the course.