Study on a New Deep Bidirectional GRU Network for Electrocardiogram Signals Classification

DOI

10.2991/iccia-19.2019.54How to use a DOI?

Keywords

ECG signals classification; Gated recurrent unit; Recurrent neural networks.

Abstract

The classification of electrocardiogram (ECG) signals has become a major issue in the medical field. And the timing characteristics of RNN are superior in the diagnosis. To cope with this problem more effectively, this paper described a new deep bidirectional gated recurrent unit (DBGRU) network. The raw input data was processed by principal component analysis (PCA) and sent to the classification model to improve performance where PCA is used for data denoising and dimensionality reduction. Several other models include unidirectional long short-term memory (ULSTM), unidirectional gated recurrent unit (UGRU), convolutional neural network (CNN) and neural network (NN) are used for comparisons. The experiment has been performed for all 23 categories of arrhythmia data obtained from the MIT-BIH arrhythmia database. The deep bidirectional GRU network was trained using the processed data and achieved a high overall accuracy of 99.51% which greatly exceeds the other four models.

Copyright

© 2019, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yun Ju
AU  - Min Zhang
AU  - Huixian Zhu
PY  - 2019/07
DA  - 2019/07
TI  - Study on a New Deep Bidirectional GRU Network for Electrocardiogram Signals Classification
BT  - Proceedings of the 3rd International Conference on Computer Engineering, Information Science & Application Technology (ICCIA 2019)
PB  - Atlantis Press
SP  - 355
EP  - 359
SN  - 2352-538X
UR  - https://doi.org/10.2991/iccia-19.2019.54
DO  - 10.2991/iccia-19.2019.54
ID  - Ju2019/07
ER  -