Scholarlink

<p>Research links the rise in adolescent drug use in South Korea to psychological stress, academic pressure, and the normalization of risky behaviors (e.g., early alcohol and tobacco use); however, the complex, nonlinear relationships among predictors remain unclear. This study used machine learning to identify the key behavioral, psychological, and physical predictors of adolescent drug use in South Korea and prevention strategies. Data from 4070 adolescents were analyzed using machine learning models, and predictors were ranked and their impact visualized. LightGBM achieved the highest accuracy (76.41%), followed by CatBoost (76.17%). Results revealed critical psychological predictors (sadness, suicidal ideation, and stress) and behavioral factors (alcohol and tobacco use). Psychological distress and risky behaviors are crucial factors in adolescent drug use, shaped by South Korea’s academic and social pressures. Conversely, physical activity and health status serve protective roles. These machine learning–driven insights can inform more effective mitigation of adolescent substance use.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Behavioral, Psychological, and Physical Predictors of Adolescent Drug Use in South Korea: Insights Obtained Using Machine Learning

  • 期刊论文

  • 发表日期：

    2025年5月5日

查看全文
  • Jun-hee Kim

摘要

Research links the rise in adolescent drug use in South Korea to psychological stress, academic pressure, and the normalization of risky behaviors (e.g., early alcohol and tobacco use); however, the complex, nonlinear relationships among predictors remain unclear. This study used machine learning to identify the key behavioral, psychological, and physical predictors of adolescent drug use in South Korea and prevention strategies. Data from 4070 adolescents were analyzed using machine learning models, and predictors were ranked and their impact visualized. LightGBM achieved the highest accuracy (76.41%), followed by CatBoost (76.17%). Results revealed critical psychological predictors (sadness, suicidal ideation, and stress) and behavioral factors (alcohol and tobacco use). Psychological distress and risky behaviors are crucial factors in adolescent drug use, shaped by South Korea’s academic and social pressures. Conversely, physical activity and health status serve protective roles. These machine learning–driven insights can inform more effective mitigation of adolescent substance use.