An empirical evaluation of benchmark machine learning classifiers for risk prediction of cardiovascular disease in diabetic males

Abstract

ABSTRACT Coronary artery disease is common in diabetics. 65% of diabetics are at risk of developing coronary artery disease or stroke, according to data from the National Heart Association from 2012. Here, classification techniques were analogized for their ability to foretell the future presence of patients with Cardiovascular Disease (CVD) in the next 10 years. For selecting important features, effective feature selection techniques like Recursive Feature Elimination (RFE) were utilized; also, for analyzing the classifiers' performance, Machine Learning (ML) classifiers were wielded. The methods like Decision Trees (DTs), K–Nearest Neighbor (KNN), Logistic Regression (LR), Artificial Neural Networks (ANNs), along with Random Forest (RF) were tested, and their outcomes were compared. To solve the classification issues, common classifiers like LR classifiers, neural networks, KNNs, RFs, and DTs are combined into a unified model. LR was predicted with the best accuracy among the predicted analysis. The study found that the LR attained the lowest rate of error with the highest accuracy (84.4%). Thus, LR is the optimal method for classification in this data set. By early prediction of coronary illness, this strategy will reduce the outstanding burden on people.