Unraveling Game Addiction Behavior Among Digital Native Students: Leveraging Machine Learning for Insights

The allure of gaming often leads to excessive engagement, which can result in addictive and problematic behaviors, negatively impacting academic performance and school achievements. While gaming becomes increasingly prevalent, concerns about game addiction have grown, particularly among students. The excessive and obsessive usage of video games that characterizes game addiction can have detrimental effects like academic deterioration, social isolation, and mental health problems. This study seeks to elucidate the complex patterns of game addiction among digital-native students by utilizing advanced machine learning techniques. We employed Random Forest, SVM, Decision Tree, and Naïve Bayes algorithms to classify game addiction data from students. To address the class imbalance, we applied Borderline-SMOTE. Our results demonstrate that decision tree and random forest outperformed other machine learning approaches, achieving an accuracy of $\mathbf{9 5 \%}$ in classifying game addiction. These findings contribute to the expanding body of knowledge on game addiction. Furthermore, in order to provide real-time monitoring and intervention for students who are at risk of developing a game addiction, future research may investigate incorporating these findings into educational platforms or applications.