<![CDATA[Academic-related stress is a critical issue that can adversely affect students’ mental health and academic performance. This study wants to create a system that can predict how stressed students are. It uses the Naïve Bayes algorithm. Naïve Bayes is a way of classifying things using probabilities. It is based on Bayes’ Theorem and makes a key assumption that each feature is independent. It works by calculating the chances of different outcomes to decide which category something belongs to. Primary data were collected from 117 students of Universitas Islam Negeri Sumatera Utara through a questionnaire based on the Transactional Model of Stress and Coping by Lazarus and Folkman, covering academic, emotional, social, and psychological factors. The research stages included data cleaning, feature selection, training-testing data splitting, model training, and performance evaluation. The results indicated an accuracy rate of 95%, with optimal performance in predicting low and moderate stress categories. The model was implemented in an Android-based application developed using Flutter and integrated with a Flask API, enabling users to perform self-assessment and receive stress management recommendations according to their predicted category. This study makes a practical contribution to the early detection of student stress levels, supporting preventive interventions and enhancing the effectiveness of mental health services in higher education institutions]]>
Copyrights © 2025
