Exploring Factors Influencing Adolescent Short Video Addiction: A Latent Profile Analysis and Machine Learning Approach

The proliferation of short-form video platforms has precipitated mounting concerns regarding addiction among adolescents. This investigation examines Short-Form Video Addiction (SFVA) through the integration of Latent Profile Analysis (LPA) and machine learning techniques. Analysis of data from 11,687 adolescents revealed four distinct user profiles: low-risk self-controlled (18.03%), high-impact low-awareness (24.60%), high-risk dependent (25.12%), and moderate-risk potential (32.25%) groups. Subsequent machine learning analysis, utilizing random forest algorithm with 89% accuracy, identified peer attachment, parental overprotection, and parenting styles as paramount predictors of SFVA development. These findings illuminate the multifaceted nature of adolescent SFVA and provide empirically-grounded recommendations for targeted intervention strategies.