Performance evaluation of different machine learning techniques for prediction of heart disease

Abstract

Heart diseases are of notable public health disquiet worldwide. Heart patients are growing speedily owing to deficient health awareness and bad consumption lifestyles. Therefore, it is essential to have a framework that can effectually recognize the prevalence of heart disease in thousands of samples instantaneously. At this juncture, the potential of six machine learning techniques was evaluated for prediction of heart disease. The recital of these methods was assessed on eight diverse classification performance indices. In addition, these methods were assessed on receiver operative characteristic curve. The highest classification accuracy of 85 % was reported using logistic regression with sensitivity and specificity of 89 and 81 %, respectively.