Classification of Stress in Students using Machine Learning Algorithms

Stress is the body’s natural reaction to change, leading to physical, emotional, and psychological feature responses. It’s supposed to assist the body cope with the situation, but it makes the mind feel like it’s on high alert. When stress becomes chronic, it can lead to health issues, including sleeplessness, depression, and heart disease, if not addressed. Exams are very stressful for students. Despite the number of notes, techniques, and study materials available for passing an exam, the most common problem is an inability to deal with exam stress. As a result, it’s vital to assist students in managing their stress levels. To fulfill this objective, in this work, we develop a model that can predict the stress level of a student. A survey was conducted where the data was collected through google forms. The Perceived Stress Scale (PSS-10) proposed by Sheldon Cohen was used to assess the students’ perceptions of stress. It’s a self-reported questionnaire that assesses how stressful a scenario is in a person’s life. Three classification algorithms, SVM, KNN, and Naive Bayes, are applied, and accuracy, precision, and recall are used as performance parameters. The highest accuracy was recorded by SVM with 95.32%.