Bibliographic Details
Title: |
Classification of inter-patient's cardiac arrhythmias in ECG signals with enhanced Jaya optimized TQWT parameters and stacked ensemble algorithm. |
Authors: |
Ramasamy, Karthikeyan1 (AUTHOR), Balakrishnan, Kiruthika1 (AUTHOR) bkiruthikaece@gmail.com, Velusamy, Durgadevi2 (AUTHOR) |
Source: |
Soft Computing - A Fusion of Foundations, Methodologies & Applications. Aug2023, Vol. 27 Issue 16, p11341-11356. 16p. |
Subject Terms: |
*ARRHYTHMIA, *FEATURE extraction, *HEART beat, *OPTIMIZATION algorithms, *ELECTROCARDIOGRAPHY, *PRINCIPAL components analysis, *WAVELET transforms |
Geographic Terms: |
SAINT Petersburg (Russia) |
Abstract: |
Cardiac arrhythmia (CA) is an irregular rhythm that can cause an increase or decrease in heart rate. Electrocardiogram (ECG) is a noninvasive diagnostic technique widely used to identify CAs. The cardiologist analyzes these ECG signals for the accurate diagnosis and treatment of CA. However, diagnosing minute changes with the naked eye for longer ECG recordings is difficult and takes much time for the cardiologist to conclude the disease type. Therefore, developing an automated diagnostic tool for classifying CA from ECG recordings is essential. In this work, we propose a novel technique for effectively diagnosing CAs from the ECG signal. The proposed method optimizes the hyper-parameters of the tunable Q wavelet transform (TQWT) and random subspace of the stacked ensemble classifier using an enhanced Jaya optimization algorithm (EJOA) for CA classification. This optimized parameter of TQWT extracts features from ECG signals by decomposing them into high- and low-pass sub-band signals. Then, the low-dimensional features are extracted from the sub-band signal coefficients by applying principal component analysis (PCA). Finally, the optimized random subspace stacked ensemble classifier is trained with the optimal random subspace features of the principal components for subsequent classification of CAs. The trained random subspace stacked ensemble classifier was independently tested with inter-patient testing beats of the MIT-BIH arrhythmia and St Petersburg INCART 12-lead arrhythmia databases to diagnose different CA classes. The proposed method achieved an average accuracy of 98.30% and a sensitivity of 93.42% for the MIT-BIH database, an average accuracy of 96.60%, and a sensitivity of 80.2% for the INCART database. These performance measures demonstrate the proposed classifier's efficacy and are better than those reported in the existing literature. [ABSTRACT FROM AUTHOR] |
|
Copyright of Soft Computing - A Fusion of Foundations, Methodologies & Applications is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) |
Database: |
Academic Search Complete |