COMPARATIVE STUDY OF CENTROID DEFUZZIFICATION AND MAXIMUM DEFUZZIFIER MODEL MEAN OF MAXIMA ALGORITHM FOR CORONARY ARTERY DISEASE DIAGNOSIS

Abstract

Objective: Cardiovascular, especially coronary artery disease (CAD), is the number one cause of death globally. Every year, more people die from cardiovascular disease than any other disease. There are several factors and symptoms of coronary artery disease that are easily recognized. To find out the severity of CAD, a diagnosis of CAD is needed to facilitate medical treatment. This study aims to explain the diagnostic process for CAD using a fuzzy system and to determine the accuracy of the fuzzy system. Design and method: This study uses a fuzzy system to diagnose the severity of coronary artery disease. Input variables used in the study were gender, age, pulse, systolic blood pressure, cholesterol, blood sugar, triglycerides, electrocardiogram (ECG), chest pain, shortness of breath, and cough. In making the system, 90 data were used which were then divided into two types of data, namely 70 training data and 20 testing data. The fuzzy inference system used in this study is the Mamdani inference system which uses centroid defuzzification and maximum defuzzifier mean of maxima (MOM). This defuzzification process is used to diagnose types of coronary artery disease, namely: CAD level 1 (Asymptomatic), CAD level 2 (Angina Pectoris), and CAD level 3 (Acute Myocardial Infarction). Results: The results of the research on the application of the fuzzy system for diagnosing coronary artery disease were the accuracy of the defuzzification centroid method of 98.5% for training data and 95% for testing data, while for the MOM defuzzification method, the accuracy level of training data was 98.5% and 90% for testing data. Conclusions: Based on the research results, it can be concluded that the centroid defuzzification method is better than the MOM defuzzification system for the coronary artery disease diagnosis system, so, it can be said that by using a fuzzy system with centroid defuzzification, the probability of correctly diagnosing one patient is 95%.