Stress-Level Classification for Students with Recommendation Outputs

Abstract

Student mental health has deteriorated measurably and continuously over the past decade. A 2025 survey by SRM University AP, published in the Asian Journal of Psychiatry, found that nearly 70 percent of 1,628 college students across eight major Indian cities reported moderate to high anxiety levels, while over 60 percent exhibited clinically significant depression symptoms. These numbers are not outliers. Cross-national data shows Japan at 80 percent stress prevalence, the United States at 67 percent, and Pakistan at 71 percent. What unites these figures is not geography but the structure of modern academic life: excessive workload, unrelenting performance pressure, financial strain, social comparison amplified by digital platforms, and institutional support systems that were simply not built for this scale of need. The clinical response to this crisis has been slow and resource-constrained. Counseling services are stretched thin. Most students who need help do not seek it, largely because of stigma, lack of awareness, or long wait times. Only 7 percent of college students in the United States access professional mental health support in any given year (EssayPro, 2025). This gap between need and access defines the problem this article addresses. This article presents a machine learning-based classification framework that predicts student stress levels across three tiers: low, medium, and high. The model draws on academic, lifestyle, physiological, and psychometric data collected through structured surveys. It evaluates five classification algorithms, including Logistic Regression, Random Forest, XGBoost, and ensemble methods, with accuracies ranging from 86.2 to 97.92 percent in peer-reviewed studies. The framework also includes a recommendation engine that generates targeted, evidence-based intervention guidance for each stress tier. The goal is to give educational institutions a practical, scalable tool for identifying students at risk before their stress reaches crisis level.