Heart disease prediction and classification using machine learning algorithms

Abstract

Heart disease is one of the most serious human diseases, with devastating consequences. Accurate and timely heart disease diagnosis is critical for heart failure prevention and treatment.Traditional medical history diagnosis of heart disease has been deemed untrustworthy in several ways. Noninvasive approaches such as machine learning are reliable and effective for classifying healthy persons and people with cardiac disease.In this research we have developed a machine learning based heart disease prediction system. We use two feature selection algorithms, two balancing techniques, four popular machine learning algorithms and one cross validation method. All of the classifiers, feature selection techniques, preprocessing methods, validation methods, and classifier performance evaluation measures utilized in this paper have been discussed.The suggested system's performance has been validated on both oversampling and undersampling and upon evaluating them we chose oversampling .The suggested system's performance has been validated on both full and reduced feature sets.The decrease of features has a benefit on classifier performance in terms of accuracy.The suggested machine-learning-based decision support system will help physicians efficiently diagnose heart patients.