Key result
A Gated Recurrent Unit model using CEEMDAN and Pearson Distance Metric correctly classified heart sound signals with an overall accuracy of 87.2%, and precision of 79% for normal and 100% for abnormal signals.
Why the study?
Subjectivity in PCG heart signal analysis when making decisions can lead to fatal misinterpretation, necessitating objective classification methods.
Can a Gated Recurrent Unit (GRU) model accurately classify normal and abnormal heart sound signals in a tele-healthcare system?
Can a Gated Recurrent Unit (GRU) model accurately classify normal and abnormal heart sound signals in a tele-healthcare system?
A GRU-based machine learning model with CEEMDAN preprocessing demonstrates high accuracy in classifying normal and abnormal heart sounds for potential tele-healthcare applications.
May support tele-healthcare phonocardiogram classification; hypothesis-generating pending prospective clinical validation.
The heart has responsibility for pumping blood throughout the body. Heart health is essential to carry out daily activities. PCG is a method of monitoring heart signals that has been used for a long time, but it is rich in information on the characteristics of heart signals. Subjectivity in Phonocardiogram (PCG) heart signal analysis in making decisions when recognizing cardiac signals is a major issue since it might lead to fatal misinterpretation. Many investigations and developments have been made to classify the PCG heart sound signal so that objective examination of the PCG heart signal may be performed. The most essential feature of our contribution is evaluating the performance of the Gated Recurrent Unit to classify heart sound signals in a tele-healthcare system using CEEMDAN (Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) and Pearson Distance Metric as preprocessing methods. The result shows CEEMDAN algorithm and Pearson Distance Metric used in the decomposition process can separate murmurs and noise from the heart sound signal, while the Shannon Energy increases the quality of the signal. Based on our evaluation, GRU correctly classifies heart sound signals, and it can be seen from the precision produced, the accuracy of 87.2%, a precision of 79% for normal heart sound and 100% for abnormal signals.
No takes yet. Share an insight, caveat, or question.
Oktarina et al. (2021) studied Heart sound signal classification. Gated Recurrent Unit (GRU) with CEEMDAN and Pearson Distance Metric was evaluated on Classification accuracy and precision. A Gated Recurrent Unit model using CEEMDAN and Pearson Distance Metric correctly classified heart sound signals with an overall accuracy of 87.2%, and precision of 79% for normal and 100% for abnormal signals.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: