Heart Disease Prediction Using a Hybrid Feature Selection and Ensemble Learning Approach

Abstract

Heart diseases have become the leading cause of death globally, highlighting the urgent need for robust diagnostic and treatment methods. This study leverages the UCI heart disease dataset to assess the effectiveness of various Machine Learning models in predicting heart diseases. This paper proposed an advanced prediction method that combines feature selection using a hybrid of Genetic Algorithm (GA) and Cuckoo Search Optimization (CSO) with a majority voting ensemble of Convolutional Neural Network and Random Forest. This approach also integrated GA for hyperparameter tuning, enhancing predictive accuracy. Comprehensive preprocessing techniques, including handling missing values, outlier detection, and normalization, were employed to ensure data quality. Using the proposed approach, 95% accuracy, 95.65% precision, 91.7% recall, 93.61% F1-score, 97.22% specificity, and 95.02% ROC AUC has been achieved. Our results demonstrate that the proposed method outperforms the existing models.