Key result
Novel LSTM deep learning model enables automated arrhythmia classification without hand-crafted feature extraction.
Why the study?
Current machine learning approaches for ECG classification rely on hand-crafted or inadequate feature learning architectures that may not optimally extract features.
Deep learning architectures, specifically LSTM networks, offer a promising approach for automated ECG arrhythmia classification by overcoming the limitations of shallow feature learning techniques.
The importance of ECG classification is on the rise with its many current medical applications of the same. Currently, there are many machine learning approaches contributed towards analysing and classifying ECG data. The main setbacks of these ML results are use of hand-crafted or inadequate feature learning architectures which often relies on the probability of finding the appropriate features. One of the proposing solutions is to use deep learning architectures where first layer uses LSTM which behave as feature extractors and summarized with some fully-connected (FCN) layers are used for predicting the final conclusion about classes in ECG
No takes yet. Share an insight, caveat, or question.
Monica.s et al. (2020) studied Arrhythmia. Deep learning with LSTM and fully-connected layers vs. Traditional machine learning techniques was evaluated. A deep learning architecture using LSTM and fully-connected layers is proposed for automated arrhythmia classification to overcome the limitations of hand-crafted feature extraction.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: