Predictive Modeling of Student Study Length Based on Academic Performance Using the Naive Bayes Algorithm

Abstract

Timely completion of study is a critical indicator of higher education quality and academic management effectiveness. Delayed graduation not only affects students but also impacts institutional performance. This study aims to develop a predictive model of student study duration based on academic performance using the Naive Bayes algorithm. The dataset consists of students’ academic records, including semester-based Grade Point Averages (GPA) from early to mid-study periods and supporting attributes. The research methodology includes data preprocessing, GPA discretization, training–testing data splitting, model construction, and performance evaluation using a confusion matrix and classification metrics such as accuracy, precision, recall, and F1-score. The results demonstrate that the Naive Bayes model achieves satisfactory predictive performance in classifying student study duration. An important finding indicates that the stability of academic performance plays a more significant role than absolute GPA values in predicting graduation timeliness. The proposed model has the potential to be implemented as an academic early warning system to identify at-risk students at an early stage and to support data-driven academic decision-making in higher education institutions.