A Hybrid SEM-ANN Approach to Investigate the Internet Addiction Among University Students Based on Psychological Resilience Theory and Cognitive-Behavioral Theory.

The internet is now essential in college students' lives, but its overuse is turning into a worldwide issue, notably with rising internet addiction among students. Earlier studies have mainly explored the risk factors of internet addiction, yielding various findings. This study aims to delve into the key factors affecting internet addiction among university students by integrating the theory of psychological resilience with cognitive-behavioral theory. It thoroughly analyzes how self-control, emotional regulation, social support, perceived stress, and psychological resilience influence internet addiction and explores their interactions and underlying mechanisms. The study conveniently selected 999 university students for a survey to measure their self-reported ratings on six constructs: self-control, emotional regulation, perceived stress, psychological resilience, social support, and internet addiction. Employing a Structural Equation Modeling - Artificial Neural Network (SEM-ANN) approach, the study unveiled complex and non-linear relationships between predictors and internet addiction. Results indicated that self-control and psychological resilience significantly reduce internet addiction, while perceived stress notably increases the risk. Notably, emotional regulation and social support did not directly lower the risk of internet addiction. Further analysis revealed that psychological resilience plays a mediating role between self-control, emotional regulation, social support, and internet addiction. Additionally, multilayer perceptron analysis of normalized importance showed self-control as the most critical predictive factor (100%), followed by emotional regulation (9.1%), social support (8.4%), and psychological resilience (5.4%). The study contributes theoretical and practical insights into internet addiction among university students.