Using Random Forest Models to Identify Predictors of Adolescent Aggression
Journal of Adolescent and Youth Psychological Studies · 2026 · 7 (2) · 1-10
DOI: 10.61838/kman.jayps.4876 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Adolescent and Youth Psychological Studies, 2026. Verfasst von Farnaz Nasiri und 2 weiteren Autor:innen.
Themen: Gewalt gegen Pflegekräfte
Abstract (Originalsprache)
Objective
The objective of this study was to identify and rank the most important individual, familial, and contextual predictors of adolescent aggression using a Random Forest machine learning approach. Methods and Materials: This study employed a cross-sectional predictive design involving adolescents aged 13–18 years recruited from public secondary schools in Tehran, Iran. Data were collected using validated self-report instruments assessing aggression and a broad range of psychological, family, peer, and school-related variables. After data screening and preprocessing, a Random Forest model was developed to predict levels of adolescent aggression. The dataset was partitioned into training and testing subsets, and model hyperparameters were optimized using cross-validation procedures. Model performance was evaluated using explained variance and error-based indices, and variable importance metrics were extracted to determine the relative contribution of each predictor.
Findings
The Random Forest model demonstrated strong predictive performance, explaining a substantial proportion of variance in adolescent aggression and maintaining good generalizability across training and testing samples. Emotion regulation difficulties emerged as the strongest predictor of aggression, followed by impulsivity and peer victimization. Self-control, parental monitoring, academic stress, and family cohesion also contributed meaningfully to the model, while demographic variables showed comparatively weaker predictive power. Incremental analyses indicated that individual psychological variables accounted for the largest increase in explained variance, with additional contributions from family and peer–school contextual factors.
Conclusion
The findings highlight the utility of Random Forest modeling for capturing the complex, multilevel determinants of adolescent aggression and demonstrate that emotional and self-regulatory processes play a central role in aggressive behavior.
Zitieren
APA
Nasiri, F., Dizaji, S. A., & Hamidi, Z. (2026). Using Random Forest Models to Identify Predictors of Adolescent Aggression. Journal of Adolescent and Youth Psychological Studies, 7(2), 1-10. https://doi.org/10.61838/kman.jayps.4876
Vancouver
Nasiri F, Dizaji SA, Hamidi Z. Using Random Forest Models to Identify Predictors of Adolescent Aggression. Journal of Adolescent and Youth Psychological Studies. 2026;7(2):1-10. doi:10.61838/kman.jayps.4876
BibTeX
@article{Nasiri2026,
author = {Nasiri, Farnaz and Dizaji, Samira Ataei and Hamidi, Zahra},
title = {Using Random Forest Models to Identify Predictors of Adolescent Aggression},
journal = {Journal of Adolescent and Youth Psychological Studies},
year = {2026},
volume = {7},
number = {2},
pages = {1--10},
doi = {10.61838/kman.jayps.4876}
}
EUVIMED