KMAN Publication Inc. Quality of Life and Health Sciences 1 Serial Number 1 2025 07 01 Machine Learning Classification of Quality of Life Risk Profiles Among Stroke Survivors Using Support Vector Machine and Decision Tree Algorithms Machine Learning Classification of Quality of Life Risk Profiles Among Stroke Survivors Using Support Vector Machine and Decision Tree Algorithms 1 15 10.61838/7h9rsg95 EN Journal Article 2025 03 17 <table> <tbody> <tr> <td> <p>Objective: This study aimed to classify quality of life risk profiles among stroke survivors in Armenia using Support Vector Machine and Decision Tree algorithms and to identify the most influential clinical, functional, psychological, and sleep-related predictors of high-risk quality of life status.</p> <p><strong>Methods and Materials:</strong> This cross-sectional predictive modeling study was conducted on 426 stroke survivors recruited from neurological, rehabilitation, and post-stroke care centers in Armenia. Data were collected using a demographic and clinical information form, the Stroke-Specific Quality of Life Scale, Barthel Index, Modified Rankin Scale, National Institutes of Health Stroke Scale, Patient Health Questionnaire-9, Generalized Anxiety Disorder-7, and Pittsburgh Sleep Quality Index. Participants were classified into high-risk, moderate-risk, and low-risk quality of life profiles. The dataset was divided into training and testing sets using stratified sampling. Support Vector Machine and Decision Tree models were trained, optimized through cross-validation, and evaluated using accuracy, precision, recall, F1-score, area under the curve, and confusion matrix indices.</p> <p><strong>Findings:</strong> Significant differences were observed among the three quality of life risk profiles in neurological severity, disability, functional independence, depressive symptoms, anxiety symptoms, sleep quality, and total quality of life scores (p &lt; 0.001). The Support Vector Machine model showed superior cross-validation performance compared with the Decision Tree model, with higher accuracy, macro F1-score, and macro AUC. In the testing set, the Support Vector Machine achieved a macro precision of 0.82, macro recall of 0.82, macro F1-score of 0.82, and macro AUC of 0.90. The Decision Tree achieved a macro precision of 0.73, macro recall of 0.74, macro F1-score of 0.74, and macro AUC of 0.81. Functional independence, global disability, depressive symptoms, sleep quality, neurological severity, and anxiety symptoms were the most important predictors.</p> <p><strong>Conclusion:</strong> Machine learning models, particularly Support Vector Machine, can effectively classify quality of life risk profiles among stroke survivors and may support individualized rehabilitation planning by identifying patients at greater risk for poor post-stroke quality of life.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/qlhs/article/download/5773/10920