Comparative Analysis of K-Means and Naïve Bayes Algorithms for Predicting Students' Academic Performance

A student's performance is a victory statistic in higher education. The university's exceptional academic record strengthens its position as one of the prerequisites for a prestigious university. Teachers need to forecast and analyze student performance to pinpoint areas of weakness and improve academic standing. In academic settings, Coordination of computational tactics to improve workforce management and academic attainment is achieved through Educational Data Mining (EDM), a theory-based approach. Classification is a broadly connected technique in forecasting student performance based on diverse criteria. Machine learning algorithms are fundamental to knowledge disclosures, permitting precise performance projection and early student-identifiable proof. This study examines how well students perform academically using the Naïve Bayes classifier (NBC) model and the K-Means clustering approach. From supervised and unsupervised machine learning, two (2) algorithms with comparable operational capacity were selected. The labeled classes in the classifier correspond to the grades in the dataset. Records were gathered from 178 students (400 levels) in Adamawa State University Mubi's computer science departments in the 2022–2023 academic sessions. The training and testing sets of the dataset are divided into two groups, each with a percentage ratio of 30% and 70%. According to the results, the Naïve Bayes model has an accuracy of 92.6%, while the K-Means model has an accuracy of 38.9%.