KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 49 2026 09 10 Classification of Adolescents With High Versus Low Emotion-Regulation Difficulties Using k-Nearest Neighbors and Psychological Indicators Classification of Adolescents With High Versus Low Emotion-Regulation Difficulties Using k-Nearest Neighbors and Psychological Indicators 1 17 EN Journal Article 2026 04 24 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> This study aimed to classify adolescents with high versus low emotion-regulation difficulties using a k-nearest neighbors (KNN) machine-learning algorithm based on self-esteem, depressive symptoms, anxiety symptoms, perceived stress, and psychological resilience.</p> <p><strong>Methods and Materials:</strong> This quantitative, cross-sectional predictive classification study was conducted among 684 adolescents aged 13–18 years enrolled in secondary schools in Ontario, Canada. Participants were classified into high and low emotion-regulation difficulty groups based on scores obtained from the Difficulties in Emotion Regulation Scale. Psychological predictors were assessed using the Rosenberg Self-Esteem Scale, Patient Health Questionnaire-9, Generalized Anxiety Disorder-7, Perceived Stress Scale-10, and Brief Resilience Scale. The dataset was divided using stratified sampling into an 80% training set and a 20% independent test set. Predictor variables were standardized before model development. Candidate KNN models with odd values of k ranging from 3 to 25 and both uniform and distance-weighted voting were evaluated using stratified 10-fold cross-validation. The final model was selected according to cross-validated F1 score and balanced accuracy and was subsequently assessed on the independent test set using accuracy, sensitivity, specificity, precision, F1 score, balanced accuracy, ROC-AUC, and confusion-matrix indices.</p> <p><strong>Findings:</strong> Adolescents with high emotion-regulation difficulties had significantly lower self-esteem and resilience and significantly higher depressive symptoms, anxiety symptoms, and perceived stress than those with low difficulties (all p &lt; .001), with large effect sizes. The optimal classifier used k = 9 with distance-weighted voting and achieved cross-validated accuracy of .855, F1 = .858, and balanced accuracy = .855. In the independent test sample, the model achieved accuracy = .861, sensitivity = .870, specificity = .853, precision = .857, F1 = .863, balanced accuracy = .861, and ROC-AUC = .923. The classifier correctly identified 60 of 69 adolescents with high difficulties and 58 of 68 adolescents with low difficulties.</p> <p><strong>Conclusion:</strong> A multivariate psychological profile comprising self-esteem, depressive symptoms, anxiety, perceived stress, and resilience can distinguish adolescents with high versus low emotion-regulation difficulties with strong out-of-sample accuracy.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jayps/article/download/6286/11946