KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 49 2026 09 10 Prediction of Depression Risk Among Adolescents Using Random Forest Based on Psychological, Family, and School Factors Prediction of Depression Risk Among Adolescents Using Random Forest Based on Psychological, Family, and School Factors 1 17 EN Journal Article 2026 04 27 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to predict elevated depression risk among Canadian adolescents using a Random Forest model incorporating psychological, family, peer, and school-related factors.</p> <p><strong>Methods and Materials:</strong> A cross-sectional predictive study was conducted among 612 adolescents aged 13–18 years recruited from secondary schools in Ontario, British Columbia, and Alberta, Canada. Depression risk was assessed using the Patient Health Questionnaire for Adolescents, with a score of 10 or higher indicating elevated risk. Psychological predictors included self-esteem, perceived stress, cognitive reappraisal, and expressive suppression; family predictors included family functioning and perceived family support; and school-related predictors included school connectedness, peer support, academic performance, attendance, and peer victimization. The dataset was divided using stratified sampling into an 80% training set and a 20% independent test set. A Random Forest classifier was optimized using stratified five-fold cross-validation and evaluated using accuracy, sensitivity, specificity, precision, F1 score, ROC-AUC, and Brier score. Predictor importance was assessed using permutation importance and SHAP values.</p> <p><strong>Findings:</strong> The optimized Random Forest model achieved a mean cross-validated ROC-AUC of .913, with accuracy of .851, sensitivity of .823, specificity of .864, and F1 score of .786. In the independent test sample, the model achieved an accuracy of .877, sensitivity of .850, specificity of .890, precision of .791, F1 score of .829, ROC-AUC of .923, and Brier score of .112. Perceived stress showed the highest predictive importance, followed by self-esteem, school connectedness, family functioning, peer victimization, perceived family support, and expressive suppression. Higher stress, peer victimization, and expressive suppression increased predicted depression risk, whereas higher self-esteem, school connectedness, family functioning, and perceived support were associated with lower predicted risk.</p> <p><strong>Conclusion:</strong> The Random Forest model demonstrated strong discrimination and stable predictive performance, indicating that adolescent depression risk can be effectively identified through the combined assessment of psychological, family, and school factors.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jayps/article/download/6251/11944