KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 49 2026 09 10 Predicting Psychological Distress Among Adolescents Using a Multilayer Perceptron Neural Network Based on Academic, Family, and Social Factors Predicting Psychological Distress Among Adolescents Using a Multilayer Perceptron Neural Network Based on Academic, Family, and Social Factors 1 17 EN Journal Article 2026 04 25 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to develop and evaluate a multilayer perceptron neural network for predicting psychological distress among Taiwanese adolescents based on academic, family, social, and selected demographic factors.</p> <p><strong>Methods and Materials:</strong> This cross-sectional predictive study was conducted among 612 adolescents aged 13–18 years enrolled in junior and senior high schools in Taiwan. Psychological distress was assessed using the Kessler Psychological Distress Scale (K10), academic stress using the Educational Stress Scale for Adolescents, family functioning using the Family APGAR Scale, and perceived social support using the Multidimensional Scale of Perceived Social Support. Academic performance, daily study time, and selected demographic variables were also assessed. The dataset was divided into training, validation, and testing subsets using a 70:15:15 ratio. A multilayer perceptron model with two hidden layers was trained using the Adam optimization algorithm, with early stopping and regularization applied to reduce overfitting. Model performance was evaluated using mean absolute error, root mean squared error, and the coefficient of determination. Predictor importance was examined using permutation-based analysis.</p> <p><strong>Findings:</strong> Psychological distress was positively associated with educational stress (r = .68, p &lt; .001), despondency (r = .64, p &lt; .001), worry about grades (r = .58, p &lt; .001), and pressure from study (r = .55, p &lt; .001), while it was negatively associated with family functioning (r = −.53, p &lt; .001), perceived social support (r = −.49, p &lt; .001), and academic performance (r = −.27, p &lt; .001). The multilayer perceptron model explained 76% of variance in the training dataset, 69% in the validation dataset, and 66% in the independent testing dataset. In the testing set, MAE was 3.67 and RMSE was 4.76. Educational stress, despondency, family functioning, worry about grades, and perceived social support were the most influential predictors.</p> <p><strong>Conclusion:</strong> Psychological distress among adolescents can be predicted with acceptable accuracy using a multilayer perceptron model incorporating academic, family, and social factors, with educational stress and family functioning emerging as especially important predictors.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jayps/article/download/6288/11948