Machine learning and feature selection-enabled optimized technique for heart disease classification and prediction

When dealing with a group of patients seeking treatment for heart-related diseases, doctorswho specialize in the diagnosis and treatment of heart-related disorders have a difficultbut critical task. It comes as no surprise that cardiovascular disease is a leading source ofmorbidity and death in contemporary society. An expert system with clear categorizationthat may assist medical professionals in identifying heart disease condition based on theclinical data of a patient is often required by physicians. The aim of this work is to providea method for the prediction and classification of cardiac disease based on machine learningand feature selection. The correlation-based feature selection (CFS) method is applied tothe input data set in order to extract relevant features for analysis. The support vectormachine with radial basis function (SVM RBF) and random forest algorithms are usedhere for data classification. Cleveland heart disease dataset is used in the experiment work.This dataset has 303 instances and 14 attributes. The accuracy, specificity and sensitivityof SVM RBF are higher than those of the random forest algorithm.