Machine Learning Enabled ECG Wavelet Analysis as a Gatekeeper for Appropriate Evaluation of Diastolic Function by Echocardiography – JASE June 2017

The prediction performance of continuous wavelet-transformed 12-lead hs-ECG (MyovistaTM, HeartSciences) with time–frequency-energy displays (Fig. 1A) for diagnosing the echocardiographic features of diastolic dysfunction was validated using machine-learning approaches with receiver-operating characteristic curves (ROC).

Partho P. Sengupta, Hemant Kulkarni, Allen J. Weiss, Negin Nezarat, Ahmad Mahmoud, Dong Li, Marilyn Graves, Ahmed Rashid, Alaa Omar, Matthew Budoff, Jagat Narula