Applying machine learning to understand the role of social-emotional skills on subjective well-being and physical health.

Abstract

Social-emotional skills are vital for individual development, yet research on which skills most effectively promote students' mental and physical health, particularly from a global perspective, remains limited. This study aims to address this gap by identifying the most important social-emotional skills using global data and machine learning approaches. Data from 61,585 students across nine countries, drawn from the OECD Social-Emotional Skills Survey, were analyzed (N_China = 7246, N_Finland = 5482, N_Colombia = 13,528, N_Canada = 7246, N_{Russia =}6434, N_Turkey = 5482, N_{South Korea} = 7246, N_Portugal=6434, and N_USA=6434). Six machine learning techniques-including Random Forest, Logistic Regression, AdaBoost, LightGBM, Artificial Neural Networks, and Support Vector Machines-were employed to identify critical social-emotional skills. The results indicated that the Random Forest algorithm performed best in the prediction models. After controlling for demographic variables, optimism, energy, and stress resistance were identified as the top three social-emotional skills contributing to both subjective well-being and physical health. Additionally, sociability and trust were found to be the fourth most important skills for well-being and physical health, respectively. These findings have significant implications for designing tailored interventions and training programs that enhance students' social-emotional skills and overall health.