The Avatar Link: Understanding Gaming Addiction with Machine Learning

Recently, a study on using machine learning to tackle gaming disorder sparked important questions. A follow-up commentary highlighted two key issues: defining the bond between users and their game characters (avatars) and improving how machine learning tools assess risks for gaming disorder. This paper aims to clear up these concepts and compare different methods for predicting gaming disorder risks. One focus is on understanding the "digital phenotype" of gaming disorder and the user-avatar bond. This bond could be crucial in understanding why some people become addicted to games. The paper also explores different ways to handle data imbalance and data splitting when using supervised machine learning techniques. By addressing these points, researchers hope to advance our understanding of gaming addiction and improve how we measure its risks.