A Machine Learning Framework for Automated Detection of Perceived Stress in Female Students

Abstract

Stress is a mental state induced by difficult circumstances. COVID-19 has severely impacted the level of stress and mental health of individuals around the globe. Pharmaceutical treatment (hypnotic or sedative drugs) has shown its ineffectiveness with several unwanted side effects to reduce mental stress. Mindfulness practices have proven their potency to reduce perceived stress (PS) and enhance satisfaction with life (SWL) without any negative consequences. This study aims to detect PS in female students and study the impact of Heartfulness meditation in reducing it. In this work, statistical exploratory data analysis (EDA) is utilized to evaluate the effectiveness of Heartfulness meditation. PS and SWL scores of Week 0 and Week 20 data are recorded from two different age groups of female students and analyzed to detect stressed students using machine learning modules. The highest PS detection accuracy 66.88% and F1-score 65.30% for group-1 are obtained using the support vector machine classifier. EDA indicates that there is a significant decrease in the PS levels and an increase in the SWL after 20 weeks of Heartfulness meditation. The results of this study indicate that the practice of Heartfulness meditation offers significant advantages to female students, regardless of age. Moreover, it was found that the practice effectively reduces stress levels among the participants.

Received: 18 November 2025 | Revised: 19 May 2026 | Accepted: 27 May 2026

Conflicts of Interest

The author declares that he has no conflicts of interest to this work.

Data Availability Statement

Data are available from the corresponding author upon reasonable request.

Author Contribution Statement

Kapil Gupta: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing, Visualization.