KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 47 2026 07 10 Hybrid XGBoost–SHAP Prediction of Adolescent Academic Achievement from Cognitive Emotion Regulation, Academic Motivation, Procrastination, and Sleep Quality Hybrid XGBoost–SHAP Prediction of Adolescent Academic Achievement from Cognitive Emotion Regulation, Academic Motivation, Procrastination, and Sleep Quality 1 13 10.61838/ EN Journal Article 2026 01 25 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> The objective of this study was to develop and evaluate a Random Forest classification model to identify adolescents at elevated risk of non-suicidal self-injury (NSSI) using trauma history, shame, impulsivity, and peer victimization as predictors.</p> <p><strong>Methods and Materials: </strong>A cross-sectional study was conducted in Chile with 1,248 adolescents aged 13–18 years (mean = 15.67, SD = 1.42) recruited from public and private schools. Participants completed validated self-report measures assessing trauma history, shame, impulsivity, peer victimization, and NSSI behaviors. The dataset was split into training (80%) and testing (20%) subsets. A Random Forest classification algorithm was applied to predict NSSI risk, and hyperparameters were optimized using grid search with five-fold cross-validation. Performance metrics included accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC). Feature importance was calculated to determine the relative contribution of each predictor variable.</p> <p><strong>Findings: </strong>Inferential analyses revealed that adolescents classified as high-risk for NSSI reported significantly higher levels of trauma history, shame, impulsivity, and peer victimization compared with low-risk peers. The Random Forest model achieved an overall accuracy of 89.2%, precision of 86.1%, recall of 84.5%, F1-score of 85.3%, and AUC of 0.938. Feature importance rankings indicated that shame was the strongest predictor, followed by trauma history, peer victimization, and impulsivity. These results suggest that emotional, developmental, interpersonal, and behavioral factors jointly contribute to adolescent self-injury risk.</p> <p><strong>Conclusion: </strong>The findings demonstrate that Random Forest algorithms can accurately classify adolescents at elevated risk for NSSI, highlighting the importance of integrating trauma history, shame, impulsivity, and peer victimization into early identification and prevention efforts. Interventions should target emotional regulation, trauma recovery, and social support to reduce the likelihood of self-injurious behaviors.</p> </td> </tr> </tbody> </table> https://www.journals.kmanpub.com/index.php/jayps/article/download/5473/11078