Artificial Neural Network-Based Prediction of Violent Behaviour Using Psychosocial and Behavioural Indicators
Real-World AI Systems · 16.10.2025 · 1 (2) · 17-32
DOI: 10.64797/rwas.v1i2.78 ↗Kurz gefasst
Fachartikel in der Zeitschrift Real-World AI Systems, online erschienen am 16.10.2025. Verfasst von Olatunde Akinrolabu.
Abstract (Originalsprache)
The increasing prevalence of violence among young people highlights the urgency of empirical, data-driven approaches capable of identifying early behavioural risk indicators. Understanding the behavioural and psychosocial factors that contribute to violent tendencies is critical for early identification and prevention. This study develops a predictive framework using an Artificial Neural Network (ANN) to classify the likelihood of violent character based on a structured survey of 1,277 respondents aged 13–30 years. The dataset incorporated key predictors, including emotional disposition, peer influence, family environment, media exposure, socioeconomic conditions, and psychological traits. After data preprocessing, the ANN was trained using a 70/30 split and further validated through 5-fold cross-validation, yielding strong and consistent performance with an average accuracy of 98.2%, precision of 96.9%, recall of 98.9%, and F1-score of 97.9%. The Receiver Operating Characteristic (ROC) analysis demonstrated excellent ability, indicating a highly reliable model. The results and findings demonstrate that violent behavioural tendencies can be reliably inferred from structured psychosocial indicators, highlighting the utility of machine learning for early detection and intervention planning. This study contributes to the body of knowledge by integrating behavioural science and machine learning to provide a practical, evidence-based tool for violence prevention initiatives.
Zitieren
APA
Akinrolabu, O. (2025). Artificial Neural Network-Based Prediction of Violent Behaviour Using Psychosocial and Behavioural Indicators. Real-World AI Systems, 1(2), 17-32. https://doi.org/10.64797/rwas.v1i2.78
Vancouver
Akinrolabu O. Artificial Neural Network-Based Prediction of Violent Behaviour Using Psychosocial and Behavioural Indicators. Real-World AI Systems. 2025;1(2):17-32. doi:10.64797/rwas.v1i2.78
BibTeX
@article{Akinrolabu2025,
author = {Akinrolabu, Olatunde},
title = {Artificial Neural Network-Based Prediction of Violent Behaviour Using Psychosocial and Behavioural Indicators},
journal = {Real-World AI Systems},
year = {2025},
volume = {1},
number = {2},
pages = {17--32},
doi = {10.64797/rwas.v1i2.78}
}
EUVIMED