Investigating Student Success Rates in Biomedical Engineering Education using Machine Learning and Descriptive Statistics

Engineering education often shows large variability in student outcomes, especially in interdisciplinary fields like biomedical engineering. The diverse curriculum, spanning physics, mathematics, anatomy, informatics, and bioethics, makes it difficult for students to achieve consistent academic performance. This study examines the effects of class attendance and exam format on student performance in a 1st-semester course offered at the Department of Biomedical Engineering, University of West Attica, Greece. Data from 2020–2024 were analyzed using descriptive statistics, t-tests, chi-square tests, monte carlo simulations and machine learning models. Results indicate attendance is critical: students attending over 80% of classes averaged a grade of 7.85 and had a 96.3% success rate, while lower attendance corresponded to a 2.74 average and 28.7% success rate. The two-part exam format also improved performance, with higher mean grades (6.90 vs. 4.59). All results were statistically significant (p < 0.001), supporting strategies promoting attendance and restructured assessments.