Machine Learning Methods in Predicting Patients with Suspected Myocardial Infarction Based on Short-Time HRV Data

Area of Science:

• Cardiology
• Biomedical Engineering
• Machine Learning

Background:

• Cardiovascular diseases are a leading cause of global mortality.
• Automated diagnostics and personalized medicine are advancing healthcare.
• Smart home integration requires advanced health monitoring.

Purpose of the Study:

• Analyze short, 10-second, 12-lead electrocardiogram (ECG) measurements.
• Classify patients with suspected myocardial infarction using machine learning.
• Identify key ECG parameters for diagnosing myocardial infarction.

Main Methods:

• Developed four machine learning models: k-nearest neighbor, radial basis function, decision tree, and random forest.
• Analyzed time-domain ECG parameters including SDNN, BPM, and IBI.
• Utilized the open PTB-XL dataset for experimental validation.

Main Results:

• The Random Forest model demonstrated superior performance with 99.63% accuracy.
• A root mean absolute error of less than 0.004 was achieved.
• Short ECG parameters effectively classify patients with suspected myocardial infarction.

Conclusions:

• Machine learning models can accurately diagnose myocardial infarction from brief ECGs.
• The Random Forest model offers a highly accurate diagnostic tool.
• This approach aids in diagnosing patients lacking other typical heart attack indicators.