Early Diagnosis and Comparative Analysis of Different Machine Learning Algorithms for Myocardial Infarction Prediction

Abstract

Heart attack alternatively known as Myocardial Infarction is one of the primary reasons of morbidity on the planet. Therefore, the diagnosis and prediction of heart disease is persuading many researchers to develop intelligent medical decision support systems. Machine Learning has been demonstrated to be viable in helping with decision making and predictions from the huge amount of clinical data delivered by the medical care. This paper aims to improve the accuracy of machine learning models which can help to make informed decision and prediction of heart attack. We have applied six machine learning classification algorithms: Support Vector Machine, Random Forest, K Nearest Neighbors, Gaussian Naive Bayes, Decision Tree and Logistic Regression. Additionally, an extensive comparison of machine learning techniques has been carried out. Our research work suggests that machine learning methods with data balancing techniques are effective tools for stroke prediction with imbalanced data. Therefore, Synthetic Minority Over-Sampling Technique (SMOTE) has been applied in our model. Hence, it is anticipated that Random Forest excels with the highest accuracy of 96% in heart attack prediction regarding performance metrics.