Predicting Student Learning Outcomes in Vocational Computer and Network Engineering Using Naïve Bayes

Abstract

This study applied the Naïve Bayes algorithm to predict student learning outcomes in the Basic Computer and Network Engineering subject at SMKN 1 Sipispis. A quantitative approach was employed, using data from 311 students, which consisted of both academic variables (assignments, midterm exams, and final exams) and non-academic variables (attendance, attitude, and learning interest). The dataset was preprocessed by cleaning, encoding, and splitting into training and testing sets using several ratios (90/10, 80/20, 70/30, and 60/40). The Naïve Bayes model was trained and evaluated using accuracy, precision, recall, and F1-score metrics. The best performance was achieved with the 80/20 data split, yielding an accuracy of 74.6%, demonstrating the model’s ability to capture probabilistic relationships between academic and non-academic factors. These findings indicate that the Naïve Bayes algorithm can effectively classify student performance levels such as Fair, Good, and Excellent, providing a reliable foundation for an automated decision support system. The developed web-based system can help teachers identify students at risk of declining performance early, enabling more adaptive and data-driven educational interventions