Ensemble Classification Model for Prediction of Heart Disease

One of the major causes of death worldwide are Cardiovascular diseases (CVDs). Elevated levels of cholesterol, blood pressure, blood glucose, and obesity are some of the parameters that can cause heart problems. Preventing premature deaths begins with detecting those who are at a high risk of CVDs and making sure that they are given adequate treatment. This paper aims to diagnose heart disease given certain health parameters. Authors have implemented an ensemble method, stacking, for classification in this paper. Different combinations of machine learning (ML) models have been used as the base-model. Logistic Regression has been used as the meta-model. Authors have used Accuracy, F14 Score, and MCC value to test and compare the models to find out which combination for the base model gives the best accuracy. Authors obtained an accuracy of 86.55% using a combination of k-Nearest Neighbors (KNN), Support Vector Machine (SVM), and Decision Trees Classification (DT), as opposed to an accuracy of 82.3%, 84.8%, and 79.8% for each of the models used individually.