Comparison between Machine Learning Algorithms for Cardiovascular Disease Prediction

Abstract

Healthcare is the cornerstone of a cohesive society and there are many diseases threaten human life. Cardiovascular diseases are one of them. Cardiovascular diseases (CVDs) are a group of heart and blood vessel problems. They include coronary heart diseases, which affect the blood vessels that supply the heart muscle; cerebrovascular illnesses, which affect the blood vessels that supply the brain; and peripheral arterial diseases, which affect the blood vessels that supply the arms and legs. In this study machine learning algorithms for adaptive perdition of cardiovascular diseases are proposed. This study aims to significantly reduce the potential failure of machine learning algorithms to predict cardiovascular diseases. This is performed by comparing seven machine learning models: Random Forest, Decision Tree, Support Vector Machine (SVM), Adaptive Boosting (Adaboost), Nave Bayes, K-Nearest Neighbors (KNN), and Logistic Regression (LR). In the proposed research, four Kaggles were selected, and the dataset relied on ten years of historical records with 13 technical features. Furthermore, seven models of machine learning (Decision Tree, Random Forest, Adaboost, SVM, Naive Bayes, KNN, and Logistic Regression) were utilized as predictors. The input values of the methods are also used to produce three different measures for evaluations.