Computational intelligence for detection of coronary artery disease with optimized features

Abstract

Coronary Artery Disease (CAD) is one of the foremost cause of mortality in almost all over the world. It falls under the category of non-communicable diseases, that are spreading at a faster pace nowadays. The factors that create a domino effect on the disease are changing life styles, unhealthy food habits, lack of exercise and other socioeconomic factors. In the past few years, with the advancement in information technology services, health sector is transformed largely and is transmitting a massive amount of medical information. With the advancement of machine learning intelligent computational methods have proved their effectiveness in almost every field. Medical field is also getting benefitted from machine learning because of its capabilities to model complex relations. This paper discusses the use of Firefly for feature subset selection with different machine learning schemes for the identification of CAD. The different techniques implemented are Random Forest, Fuzzy Unordered Rule Induction, Logistic regression and Multilayer perceptron using Keras. Deep learning based method outperforms other learning schemes with the accuracy of 89.77%. Thus, the method can pose as a promising tool for screening CAD patients more accurately.