SVM-DEA-CSA: An innovative prognostic model for mortality prediction in heart failure utilizing metaheuristic algorithms

Cardiovascular disease remains one of the leading causes of global morbidity and mortality due to its complex pathophysiology and substantial impact on public health. Predicting outcomes in patients with heart failure is particularly challenging because clinical data are often heterogeneous, nonlinear, and influenced by multiple interacting variables. To address these challenges, this study proposes the support vector machine - differential evolution algorithm - crow search algorithm (SVM-DEA-CSA) approach. This innovative hybrid model integrates Support Vector Machine (SVM) classification with differential evolution algorithm (DEA) for feature selection and crow search algorithm (CSA) for optimal parameter tuning. Utilizing the UCI (Heart Failure Clinical) dataset for mortality prediction. The model utilizes metaheuristic optimization to enhance predictive performance and address the nonlinear and complex nature of heart failure patient data. This enhances both computational efficiency and the interpretability of the predictive model. These results have improved global search capability and superior optimization performance. The effectiveness of the proposed SVM-DEA-CSA framework is thoroughly evaluated using performance indicators, including accuracy, sensitivity, specificity, F1-score, and mean squared error. The SVM-DEA-CSA model achieved the highest score of Accuracy of 93%, Precision 100%, Sensitivity 79%, and F1-score 88% based on predictive accuracy and reliability. Overall, this framework offers a robust and efficient tool for early prognosis of heart failure, supporting timely clinical decision-making. Policymakers should improve the digital healthcare system and incorporate AI-powered predictive models.