A Hybrid Machine Learning Model to Predict Heart Disease Accurately

Abstract

Objective: To propose the most effective machine learning algorithm for predicting cardiac problems. Methods: The dataset used for this study is “heart” which was taken from www.kaggle.com. The heart dataset contains 13 features and a target variable. It is divided into 70 percent training set and 30 percent testing set. K-Fold cross-validation is used in this study for model evaluation and model selection. The K value chosen is ten. A Hybrid Ensemble machine learning model is built using a heterogeneous collection of weak learners in this work. To construct a hybrid ensemble model, weak learners such as “Logistic Regression”, “Decision Tree”, “Support Vector Machine”, “KNearest Neighbor”, and “Naive Bayes” are used. Normally, in an ensemble model, a homogeneous group of weak learners is utilized, however in this study, a heterogeneous group of weak learners is used. The parameter used in this study is accuracy. Accuracy of all the weak learners is found and compared with the hybrid ensemble model. Findings: Weak machine learning models are combined to create an ensemble model. The ”Hybrid Ensemble model” has a 98 percent accuracy rate and outperforms all weak learners such as “Logistic Regression”, “Decision Tree”, “Support Vector Machine”, “K-Nearest Neighbor”, and “Naive Bayes”. Novelty and applications : For the prediction of heart problems, the hybrid ensemble model is recommended since it extracts more accurate and valuable data from huge amounts of data, making prediction easier for physicians. Keywords: Machine Learning; Ensemble Model; Weak Learners; Homogeneous; Heterogeneous