Table 6 Best performing feature subsets for each ML model and their selection algorithm (with expert verdict).

From: Classification models for assessing coronary artery disease instances using clinical and biometric data: an explainable man-in-the-loop approach

ML

Subset

Algorithm

SVM

['known CAD', 'previous PCI', 'previous CABG', 'Diabetes', 'Smoking', 'Dyslipidemia', 'Angiopathy', 'Chronic Kidney Disease', 'ASYMPTOMATIC', 'ATYPICAL SYMPTOMS', 'ANGINA LIKE', 'RST ECG', 'male', '40–50', 'Doctor: Healthy']

GA

DT

['previous AMI', 'previous CABG', 'Arterial Hypertension', 'Angiopathy', 'Chronic Kidney Disease', 'Family History of CAD', 'ANGINA LIKE', 'male', ' < 40', 'Doctor']

GA

KNN

['known CAD', 'previous AMI', 'previous CABG', 'Diabetes', 'Smoking', 'Arterial Hypertension', 'Dyslipidemia', 'Angiopathy', 'Chronic Kidney Disease', 'ASYMPTOMATIC', 'ATYPICAL SYMPTOMS', 'ANGINA LIKE', 'INCIDENT OF PRECORDIAL PAIN', 'male', 'Overweight', ' < 40', '50–60', 'Doctor']

GA

ADA

['known CAD', 'previous AMI', 'Diabetes', 'Family History of CAD', 'ATYPICAL SYMPTOMS', 'ANGINA LIKE', 'INCIDENT OF PRECORDIAL PAIN', 'RST ECG', 'male', 'Overweight', 'Obese', ' < 40', 'Doctor: Healthy']

GA

RF

['known CAD', 'previous PCI', 'Diabetes', 'Chronic Kidney Disease', 'ANGINA LIKE', 'RST ECG', 'male', ' < 40', 'Doctor']

GA

Back to article page