<![CDATA[Cardiovascular disease (CVD) remains the leading cause of death worldwide, emphasizing the need for accurate and reliable diagnostic models. Support Vector Machine (SVM) has demonstrated strong performance in medical data classification due to its ability to handle complex and high-dimensional data; however, its effectiveness depends heavily on appropriate hyperparameter selection. To address this limitation, this study integrates the Nutcracker Optimization Algorithm (NOA), a population-based metaheuristic inspired by squirrel foraging behavior, to optimize SVM hyperparameters for cardiovascular disease prediction. Using a standardized heart disease dataset, three models were evaluated: a baseline SVM, an NOA-optimized model, and an integrated SVM–NOA model. The proposed SVM–NOA approach achieved the best performance, improving accuracy from 83.61% to 88.52%, precision from 78.12% to 86.21%, and F1-score from 83.33% to 87.72%, while maintaining a recall of 89.29%. Although NOA incurs additional one-time optimization cost, the final optimized SVM required only 0.0095 s for training, compared to 0.0205 s for the baseline SVM. These results demonstrate that SVM–NOA provides a robust and computationally practical approach for enhancing cardiovascular disease prediction.]]>
Copyrights © 2026
