Utilizing voting classifiers for enhanced analysis and diagnosis of cardiac conditions

Abstract

Identifying heart disease based on initial symptoms poses a considerable difficulty in the modern era. Untimely diagnosis may lead to fatality. An accurate decision support system is essential for timely identification of heart diseases. The model proposed is named Heart Disease Prediction Model (HDPM) and comprises three primary components; which are; (i) data collection and preprocessing, (ii) feature selection, and (iii) Disease Prediction. In the first part, the used heart disease dataset is preprocessed and the heart disease features are extracted. Then, these extracted features are fed to the second part (i.e. feature selection). This paper presents a novel approach to feature selection using the Sand Cat Swarm Optimization (SCSO) algorithm. An enhanced methodology has been implemented in the SCSO system to improve its effectiveness in identifying and categorizing the most crucial and impactful features for predicting and classifying patients with heart disease. The proposed methodology is called Dynamic SCSO (DSCSO). DSCSO is combination method between SCSO and Dynamic Opposite Learning (DOL). Ultimately, the chosen features are inputted into the voting classifiers to arrive at the ultimate determination. The proposed voting classifiers is based on using multiple classifiers which are, Logistic Regression (LR), Naïve Bayes (NB), Random Forest (RF), Extreme Gradient Boost (EGB), Decision Tree (DT), and Support Vector Machine (SVM). At the end, the proposed model (i.e., HDPM) trained and tested using heart disease data and it performs well.