Detecting Deviant Behavior Among Youth: A Socio- Psychological and Machine Learning Prognostic Model for Addiction, Suicide, and Radicalization Risks

Abstract

This study addresses one of the most pressing global challenges — the rise of deviant behavior among youth, manifesting as drug addiction, suicidal tendencies, and radicalization. The research integrates sociological and psychological analysis with a data-driven predictive model using synthetic datasets and machine learning algorithms (Random Forest and XGBoost). A dataset of 5,000 respondents was simulated to classify individuals into four risk levels (low, moderate, high, critical) based on responses to 20 behavioral and psychological indicators. The models achieved strong accuracy in identifying high-risk individuals and key behavioral predictors. The results demonstrate the potential of combining socio-psychological understanding with artificial intelligence tools for early detection and prevention of deviant behavioral tendencies among young people.