KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 48 2026 08 10 Classifying Panic-Symptom Risk among Taiwanese Youth Based on Anxiety Sensitivity, Interoceptive Awareness, Sleep Disturbance, and Perceived Stress Using a Support Vector Machine Classifying Panic-Symptom Risk among Taiwanese Youth Based on Anxiety Sensitivity, Interoceptive Awareness, Sleep Disturbance, and Perceived Stress Using a Support Vector Machine 1 19 EN Journal Article 2025 12 13 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to develop and internally validate a support vector machine model for classifying elevated panic-symptom risk among Taiwanese youth based on anxiety sensitivity, multidimensional interoceptive awareness, sleep disturbance, and perceived stress.</p> <p><strong>Methods and Materials: </strong>A cross-sectional predictive classification study was conducted with 512 Taiwanese youths aged 15–24 years who were recruited from senior high schools, vocational schools, universities, and community youth centers. Participants completed the Anxiety Sensitivity Index–3, Multidimensional Assessment of Interoceptive Awareness–2, Pittsburgh Sleep Quality Index, Perceived Stress Scale–10, and Panic Disorder Severity Scale–Self-Report. Panic-symptom risk was categorized using a predetermined cutoff score. The sample was stratified into a training set of 410 participants and an independent testing set of 102 participants. A radial basis function support vector machine was developed using repeated stratified 10-fold cross-validation, class weighting, probability calibration, and hyperparameter optimization. Model performance was assessed using accuracy, balanced accuracy, sensitivity, specificity, predictive values, F1 score, receiver operating characteristic area under the curve, precision–recall area under the curve, and Brier score.</p> <p><strong>Findings: </strong>All psychological predictors differed significantly between elevated-risk and low-risk participants, with the largest effects observed for anxiety sensitivity, perceived stress, body trusting, interoceptive not-worrying, and sleep disturbance, all p &lt; .001. In the independent testing dataset, the support vector machine achieved an accuracy of .882, balanced accuracy of .875, sensitivity of .857, specificity of .892, positive predictive value of .750, negative predictive value of .943, and F1 score of .800. The model produced an area under the receiver operating characteristic curve of .928 and an area under the precision–recall curve of .836. Anxiety sensitivity was the most influential predictor, followed by perceived stress, body trusting, sleep disturbance, interoceptive not-worrying, and interoceptive self-regulation.</p> <p><strong>Conclusion: </strong>Panic-symptom risk among Taiwanese youth can be classified with high accuracy using a multidimensional psychological profile, particularly anxiety sensitivity, perceived stress, body trust, sleep disturbance, and interoceptive regulation.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jayps/article/download/5687/11791