Implementation of the Naive Bayes Algorithm for Classifying Students’ Stress Levels Based on Learning Behavior Patterns

Abstract

This study aims to implement the Gaussian Naive Bayes algorithm to classify students’ stress levels based on learning behavior patterns as a machine learning-based early detection approach. The study employs an empirical experimental method using the “Student Stress and Study Behavior” dataset, which includes attributes such as study duration, sleep hours, academic workload, physical activity, social media use, and stress levels as the target variable. The research stages included data preprocessing, attribute transformation, dataset splitting using the hold-out validation method, classification model development, and evaluation using accuracy, precision, recall, and F1-score. The results show that the Gaussian Naive Bayes model achieved an accuracy of 91.27%, a precision of 90.84%, a recall of 91.02%, and an F1-score of 90.93%. The findings indicate that sleep duration and study duration are attributes that significantly contribute to the classification of students’ stress levels. This study contributes to the development of an efficient, interpretable, and practical decision support system to support the monitoring of students’ mental health.