High-Accuracy Heart Disease Prediction Using Bio-Inspired Optimization

Abstract

Heart disease remains one of the leading causes of death worldwide, making accurate and early diagnosis essential in modern healthcare. This paper proposes a hybrid predictive model that combines the Moth-Flame Algorithm (MFA) for feature selection with a Radial Basis Function Support Vector Machine (RBF-SVM) classifier to improve heart disease diagnosis. MFA eliminates redundant and irrelevant features and optimizes the SVM hyperparameters, thereby enhancing model efficiency and accuracy. The proposed hybrid MFA-RBF-SVM model was evaluated on three benchmark datasets: Cleveland, Hungarian, and Statlog. Experimental results show that the model outperforms conventional classifiers such as Decision Tree, Random Forest, K-Nearest Neighbor, Artificial Neural Network, and standard SVM, achieving up to 95.2% accuracy, 0.96 recall, and a 0.93 F1-score on the Hungarian and Cleveland datasets, with low false-positive and false-negative rates. These findings demonstrate the effectiveness of combining machine learning with bio-inspired optimization for reliable, adaptable clinical prediction.