Abstract
Heart disease is a type of chronic disease that can lead to death if not diagnosed in time. A clinical decision support system developed using machine learning technology can be used in the diagnosis of disease. Proper utilization of feature selection increases classification accuracy by reducing the computational cost. The purpose of this research work is to purpose a new Hybrid Pearson Correlation with Backward Elimination (HPCBE) feature selection method. Proposed HPCBE method is further used to develop a hybrid system for heart disease diagnosis (HSHDD).HPCBE is proposed by combining pearson correlation (PC) and backward elimination (BE) methods. Reduced feature subset selected by HPCBE method is used along with decision tree (DT), k-nearest neighbor (KNN), extreme gradient boosting (XGBoost) and adaptive boosting (AdaBoost) classifiers to develop HSHDD. The feature reduction ratio of 53.84% is achieved by the proposed HPCBE Feature Selection method. HSHDD achieved a maximum accuracy of 86.49% in heart disease classification with AdaBoost classifier.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have