Early Detection Of Cardiovascular Disease Using Multi-Classifier Machine Learning Approaches

Abstract

In order to increase survival rates and get treatment faster, early and accurate diagnostic systems are needed for cardiovascular disease (CVD), which is still a big worldwide health concern.  In order to improve early identification of CVD, this study suggests an advanced ML framework that integrates various classification methods with a meta-heuristic feature selection method.  The model uses an ensemble technique that combines K-Nearest Neighbor (KNN), Naive Bayes, and Support Vector Machine (SVM) classifiers. This allows the model to take use of each classifier's capabilities while reducing susceptibility to noisy data and overfitting.  By using the Imperialist Competitive Algorithm (ICA) for feature selection, the study optimizes prediction accuracy by reducing data dimensionality while keeping crucial clinical information.  A thorough evaluation of the model was guaranteed by preprocessing the medical dataset and dividing it in half for testing and training.  The results show that when compared to individual classifiers, the proposed ICA-based multi-classifier ensemble obtains better accuracy, precision, recall, and F1-score.  Cardiovascular decision support systems benefit from this hybrid approach because it increases diagnostic accuracy while also making them more trustworthy and easier to understand.