Machine Learning Predictive Model for Analyzing the Influence of Academic Performance on Course Completion in Algorithms and Programming

The success of students in core computer science courses such as Algorithms and Programming is a critical factor in their academic journey, as it reflects both mastery of fundamental concepts and readiness for more advanced studies. Academic performance in this course is not only shaped by grades but also by behavioral and psychological attributes that influence learning outcomes. This study investigates the influence of academic performance on graduation in Algorithms and Programming using a predictive machine learning approach. The dataset includes 106 student records encompassing academic variables (attendance, average grades, assignment scores), psychological factors (motivation, anxiety toward examinations), and behavioral indicators (discussion participation, AI tool usage, online learning activities). The research adopts the SEMMA methodology, consisting of sampling, exploration, modification, modeling, and assessment. Several classification algorithms were tested, and Random Forest was selected as the primary model due to its strong performance and interpretability. The results indicate that academic achievement variables, particularly average grades and attendance, significantly influence graduation. Additionally, non-academic factors such as motivation, discussion activity, and exam anxiety contribute to predictive outcomes. The model achieved an accuracy of around 91% and an AUC score of 0.93, confirming its reliability in distinguishing between students who passed and those who did not. These findings highlight that academic performance influences success in algorithm and programming courses.