<![CDATA[The expansion of digital medical records and clinical data has strengthened the development of intelligent analytical systems to support early disease detection and improve diagnostic accuracy. This study aims to evaluate the performance of three classification algorithms, namely Random Forest, Support Vector Machine, and Logistic Regression, in predicting stroke risk using multidimensional patient clinical information. The dataset consists of 224 patient records derived from the Kaggle Stroke Dataset and additional questionnaire data collected from hospitals and primary health centers. The variables include demographic characteristics, clinical history, lifestyle factors, and physiological indicators. The research methodology involves several stages, including data preprocessing, feature selection using ANOVA F value, class balancing through the Synthetic Minority Oversampling Technique, model training, and performance evaluation using Accuracy, Precision, Recall, F1 Score, Matthews Correlation Coefficient, and Area Under the Curve. The results indicate that the Random Forest model achieves the highest performance, with an accuracy of 0.91 and an Area Under the Curve of 0.91, outperforming Support Vector Machine and Logistic Regression. This outcome confirms the effectiveness of ensemble based approaches in identifying complex nonlinear patterns and managing imbalanced data. The study contributes to healthcare quality improvement by providing a reliable prediction framework that supports early clinical decision making, reduces diagnostic delays, and enhances patient care outcomes.]]>
Copyrights © 2025
