Early Detection of Myocardial Infarction Using Machine Learning with Maximum Accuracy

Abstract

This paper presents a technique for detecting Myocardial Infarction (MI) using Machine Learning through the ECG data. In today’s scenario, MI is one of the major causes of demise worldwide. MI occurs due to coronary heart disease and if detection or treatment is not done at appropriate time, the untreated MI may present with serious late complications. For treating MI in earlier stage several methods have been employed, but the parameters employed for those does not provide as much accuracy. So, a system which detects the abnormality with maximum accuracy is proposed. For that, Pan Tompkins algorithm is employed to filter and remove the noise in the acquired ECG signal and S-T segment is extracted from it. The features that are obtained from the extraction of S-T segment when given to the classifier provided better results. Based on the application of the Machine Learning classifiers, Naïve Bayes and Decision tree, the Decision tree classifier gave a higher accuracy of 98.5% compared to Naïve Bayes of 93.9%.