KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 45 2026 05 01 A Deep Neural Network Analysis of Youth Psychological Distress Based on Cybervictimization, Loneliness, Academic Pressure, Screen-Time Fragmentation, and Family Emotional Climate A Deep Neural Network Analysis of Youth Psychological Distress Based on Cybervictimization, Loneliness, Academic Pressure, Screen-Time Fragmentation, and Family Emotional Climate 1 13 10.61838/kman.jayps.5410 EN Journal Article 2025 09 13 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> The present study aimed to investigate youth psychological distress through a deep neural network framework based on cybervictimization, loneliness, academic pressure, screen-time fragmentation, and family emotional climate among adolescents and emerging adults in Georgia.</p> <p><strong>Methods and Materials: </strong>This study employed a quantitative cross-sectional predictive design using deep learning methodologies. The statistical population consisted of adolescents and emerging adults aged 16 to 24 years enrolled in secondary schools and universities in Georgia during the 2025–2026 academic year. Using multistage cluster sampling, 842 participants were initially recruited, and after data screening, 814 cases were retained for final analysis. Data were collected using standardized instruments including the Kessler Psychological Distress Scale, Cyberbullying Victimization Scale, UCLA Loneliness Scale, Educational Stress Scale for Adolescents, a screen-time fragmentation measure, and the Family Environment Scale. Data analysis involved descriptive statistics, Pearson correlation coefficients, and deep neural network modeling using Python, TensorFlow, and Keras libraries. Model performance was evaluated using mean squared error, root mean squared error, mean absolute error, coefficient of determination, and cross-validation procedures. SHAP analysis was additionally employed to examine relative predictor importance.</p> <p><strong>Findings: </strong>The findings demonstrated significant positive relationships between psychological distress and cybervictimization, loneliness, academic pressure, and screen-time fragmentation, while family emotional climate showed a significant negative relationship with psychological distress. Loneliness emerged as the strongest predictor within the deep neural network model, followed by family emotional climate and cybervictimization.</p> <p><strong>Conclusion:</strong> Comparative analyses indicated that the deep neural network outperformed traditional machine learning approaches including random forest, support vector machine, and multiple linear regression models in predicting psychological distress outcomes.</p> </td> </tr> </tbody> </table> https://www.journals.kmanpub.com/index.php/jayps/article/download/5410/9855