Quantitative analysis of socioeconomic determinants on student mental health: a machine learning approach with policy implications

This study examines the relationship between socioeconomic factors and mental health outcomes among 10,000 students using a comprehensive dataset encompassing academic performance, family income, social interactions, and psychological metrics. Findings reveal that students from low-income families experience 23% higher stress levels and 18% lower mental health scores compared to their high-income peers, while social support systems and counseling services reduce severe stress incidence by 34%. A logistic regression model achieved 82.4% accuracy (SD ±1.2%) in classifying stress levels, with academic performance (β = −0.45, p < 0.001), sleep duration (β = −0.32, p < 0.01), and family income (β = −0.28, p < 0.05) emerging as dominant predictors. Additionally, machine learning models such as XGBoost and SVM outperformed logistic regression, both achieving 99.8% accuracy, while Random Forest exhibited potential overfitting with 100% accuracy. These results highlight the need for institutional interventions, including income-based scholarships to alleviate financial burdens, structured sleep hygiene programs to promote student well-being, and mobile-based counseling services to enhance mental health resource accessibility. By implementing data-driven strategies, educational institutions can effectively address socioeconomic disparities and support student mental health and academic success.