KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 45 2026 05 01 Random Forest Prediction of Adolescent Risk-Taking Behaviors Based on Sensation Seeking, Online Disinhibition, Emotional Impulsivity, Peer Deviance, and Executive Dysfunction Random Forest Prediction of Adolescent Risk-Taking Behaviors Based on Sensation Seeking, Online Disinhibition, Emotional Impulsivity, Peer Deviance, and Executive Dysfunction 1 14 10.61838/kman.jayps.5479 EN Journal Article 2025 09 15 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> The present study aimed to predict adolescent risk-taking behaviors using a Random Forest machine learning model based on sensation seeking, online disinhibition, emotional impulsivity, peer deviance, and executive dysfunction among adolescents in Argentina.</p> <p><strong>Methods and Materials: </strong>The study employed a cross-sectional predictive-correlational design with a machine learning framework. The statistical population consisted of Argentine adolescents enrolled in secondary schools in Buenos Aires, Córdoba, and Rosario during the 2025–2026 academic year. Using multistage cluster sampling, 1,284 adolescents aged 14 to 18 years were selected. Data were collected using the Adolescent Risk-Taking Questionnaire, Brief Sensation Seeking Scale, Online Disinhibition Scale, UPPS-P Negative Urgency Subscale, Peer Deviance Scale, and the Behavior Rating Inventory of Executive Function. Preliminary statistical analyses were conducted using SPSS-28, while the Random Forest predictive model was implemented in Python using the Scikit-learn library. Data were divided into training and testing subsets using an 80/20 split procedure. Model performance was evaluated through accuracy, precision, recall, F1-score, area under the curve (AUC), cross-validation scores, and SHAP feature-importance analyses.</p> <p><strong>Findings: </strong>The findings demonstrated significant positive correlations among all study variables. Peer deviance showed the strongest positive association with adolescent risk-taking behaviors, followed by emotional impulsivity and executive dysfunction. The Random Forest model demonstrated strong predictive performance, with testing accuracy of 0.87 and AUC of 0.90, indicating excellent classification capability. Feature importance analysis revealed that peer deviance was the most influential predictor, followed respectively by emotional impulsivity, executive dysfunction, online disinhibition, and sensation seeking. SHAP analyses further indicated substantial interaction effects among emotional, cognitive, and social variables in predicting adolescent risk-taking behaviors.</p> <p><strong>Conclusion: </strong>The findings suggest that adolescent risk-taking behaviors emerge through complex interactions among emotional dysregulation, cognitive-control deficits, peer-related influences, and digital behavioral tendencies. The strong performance of the Random Forest model highlights the usefulness of machine learning approaches for identifying high-risk adolescents and understanding multidimensional psychosocial predictors of maladaptive behaviors. Early prevention efforts targeting emotional impulsivity, executive dysfunction, peer deviance, and online disinhibition may contribute significantly to reducing behavioral risk among adolescents.</p> </td> </tr> </tbody> </table> https://www.journals.kmanpub.com/index.php/jayps/article/download/5479/9883