Machine Learning Model for Student Stress Level Prediction

Stress amongst students is a very common problem in today’s society that affects both the academic achievement of students as well as their health. Mental well-being and efficiency in the performance of academic tasks can be improved by appropriate handling of stress. This study examines the application of machine learning algorithms, such as Naive Bayes (NB) and K-Nearest Neighbors (KNN), to classify stress levels among students. The classification is based on responses to questionnaires derived from the Perceived Stress Scale (PSS) and the stress component of the DASS-21 (Depression, Anxiety, and Stress Scale). These tools are utilized to detect stress-related symptoms and analyze the effectiveness of the algorithms in stress level prediction. In this study determines the classification performance model is determined by the F1-score, recall, accuracy, and precision. The conclusion made from the comparative study indicates that Naive Bayes did better than the KNN algorithms in the aspect of reliability with an accuracy of 90.0% and 77.08% in both datasets, respectively. When stress is detected early, institutions can give them the support they need, and this will result in enhancing their well-being and academic performance.