Using Machine Learning to Identify Predictors of Juvenile Drug Use and Delinquency

• In Event: Drug Treatment, Program Effectiveness, and Predictive Analytics

Abstract

Juvenile drug use is a significant public health issue with long-term consequences for individuals and society. It is often intertwined with other risky behaviors, including smoking, alcohol consumption, early sexual activity, and violence. The age of onset of drug use critically influences the likelihood of addiction. Understanding youth drug use within the broader framework of juvenile delinquency provides a more comprehensive perspective on its causes and consequences. This shift in perspective not only informs this study’s analytical approach but also has implications for how society discusses and addresses adolescent substance use. This study will analyze Youth Risk Behavior Survey (YRBS) data from 1991 to 2023 to identify key predictors of adolescent drug use using artificial intelligence (AI) and machine learning (ML) techniques. Prior research has shown ML’s effectiveness in analyzing large-scale health datasets to predict behaviors. By applying algorithms such as decision trees, random forests, and neural networks, this study will identify complex relationships among risky behaviors, substance use, and demographic factors. This approach aims to improve predictive accuracy and uncover hidden patterns in juvenile delinquency and drug use. The findings will enhance risk factor identification and inform data-driven prevention strategies for at-risk youth.

Authors

• Hyesun Kim, Worcester State University

• Hyoungah Park, Saint Peter's University

• Byung Jun Cho, Westfield State University

• Seokbeom Kim, Alfred State College