Diagnosis of cardiac diseases using spectral features and machine learning

This study presents combination of spectral analysis method, generalized discriminant analysis (GDA), and the extreme learning machine (LELM) classifier algorithms which provide a robust method for prediction of cardiac diseases. Firstly, spectral approaches, has been applied for features extraction from heart rate variability (HRV) signal. The proposed GDA +ELM model was fitted with the features derived from approaches for the classification model of congestive heart disease and coronary artery disease. The standard database of HRV signal was collected from Physio net ATM's HRV source for training and validating predictive models for cardiac diseases.