Utilising Machine Learning to Predict Myocardial Infarction by Electrocardiogram Derived Respiration

Abstract

Myocardial Infarction (MI) is one of the leading causes of death. Electrocardiogram (ECG) is a non-invasive tool that is commonly used as a diagnostic tool to assess cardiac conditions. A dataset consisting ECG signals of healthy individuals and MI patients was subjected to pre-processing techniques like normalization and application of a bandpass filter. R-R peak intervals from the pre-processed ECG signals are extracted to generate the respiratory signal. The features extracted from the respiratory signal are used to predict MI. The objective of this study is to evaluate the potential of ECG derived respiratory signal (EDR) in predicting MI by utilizing machine learning techniques like random forest, linear regression, Convolutional Neural Network(CNN), Multilayer perceptron(MLP). The results of the study will examine the feasibility of using EDR in predicting MI and provide insight into the most effective machine learning technique for this application. This study will contribute to the development of new and efficient prediction methods for MI patients.