Open Access

Introducing the hybrid “K-means, RLS” learning for the RBF network in obstructive apnea disease detection using Dual-tree complex wavelet transform based features

Javad Ostadieh
Javad Ostadieh
 and 
Mehdi Chehel Amirani
Mehdi Chehel Amirani
   | Mar 18, 2020
Journal of Electrical Bioimpedance's Cover Image
About this article
Previous Article
Next Article
Cite
Published Online: Mar 18, 2020
Page range: 4 - 11
Received: Dec 27, 2019
DOI: https://doi.org/10.2478/joeb-2020-0002
Keywords
© 2020 Javad Ostadieh et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.

Fig. 1

The overall steps of the OSA detection with the help of ECG signals.
The overall steps of the OSA detection with the help of ECG signals.

Fig. 2

The proposed OSA detection method.
The proposed OSA detection method.

Fig. 3

The first 3 seconds of the apnea ECG from an example record.
The first 3 seconds of the apnea ECG from an example record.

Fig. 4

The three level dual-tree complex wavelet transform.
The three level dual-tree complex wavelet transform.

Fig. 5

The sub bands of the ECG signal for Tree A.
The sub bands of the ECG signal for Tree A.

Fig. 6

The sub bands of the ECG signal for Tree B.
The sub bands of the ECG signal for Tree B.

Fig. 7

The absolute energy of the sub band signal x000a.
The absolute energy of the sub band signal x000a.

Fig. 8

The absolute energy of the sub band signal x000b.
The absolute energy of the sub band signal x000b.

Fig. 9

The proposed hybrid RBF classifier.
The proposed hybrid RBF classifier.

List of the used abbreviations.

Abbreviations Descriptions
OSA Obstructive sleep apnea
ECG Electrocardiogram
EDR ECG-Derived Respiration
AHI Apnea-Hypopnea Index
HMM Hidden Markov model
RUSBoost Random under-sampling Boost
Adaboost Adaptive boost
DWT Discrete wavelet transform
TQWT Tunable Q-factor wavelet transform
LDA/QDA Linear/Quadratic Discriminant Analysis
SFFS Sequential forward feature selection
SRDA Spectral regression discriminant analysis
DNN/CNN Deep/Convolutional neural network
DT classifier Decision tree classifier
RBF Radial basis function
SVM Support vector machine
RLS Recursive least squares
GS Gram-Schmidt
STLF Short-time load forecasting
DT-CWT Dual-tree complex wavelet transform

The comparison of the OSA detection results based on various methods.

References Feature extraction/s election method Classifier Results
ACC% Sens% Spec%
[1] Zarei 2018 DWT+SFFS SVM (RBF kernel) 92.98 91.74 93.75
[2] Song 2016 HMM HMM+SVM 86.2 82.6 88.4
[3] Hassan 2017 TQWT RUSBoost 88.88 87.58 91.49
[4] Gonzalez 2017 Cepstrum+ Filter bank QDA 84.76 81.45 86.82
[5] Hassan 2016 Statistical and spectral Bootstrap aggregating 85.97 84.14 86.83
[6] Hassan 2016 Normal invers Gaussian modeling AdaBoost 87.33 81.99 90.72
[7] Sharma 2016 QRS features LS-SVM (RBF kernel) 83.8 79.5 88.4
[8] Hilmisson 2018 Frequency features Statistical analysis 93 100 81
[9] Janbakhshi 2018 Time domain feaures+PSD SVM-KNN-NN-LD-QD 90.9 89.6 91.8
[10] Ma 2019 Statistical features Statistical analysis 87 89 79
[11] Nishad 2018 Tunable-Q wavelet transform features Random Forest 92.78 93.91 90.95
[12] Wang 2019 RR-intervals CNN (LeNet-5) 92.3 90.9 100
[13] Singh 2019 Time-frequency Scalogram features CNN (AlexNet) 86.22 90 100
[14] Urtnasan 2018 RR-intervals CNN 96 96 96
[15] Wang 2018 RR-intervals CNN 97.8 100 93
[16] Sharma 2019 Fuzzy-entropy (FUEN) and the Log of signal-energy (LOEN) KNN-DT-SVM 90.87 92.43 88.33
[17] Avci 2015 DWT+PCA Random forest 92–98 - -
[18] Rachim 2014 DWT+PCA SVM 94.3 92.65 92.2
Proposed method DT-CWT+SRDA Hybrid “k-means, RLS” RBF 95.62 96.37 96

List of non-linear features that are extracted from the DT-CWT coefficients in this paper.

Features Description
FE Fuzzy Entropy
ApEn Approximate Entropy
IQR Interquartile Range
RP Recurrence Plot
SD1, SD2, SD1/SD2 Poincare Plot
eISSN:
1891-5469
Language:
English
Publication timeframe:
Volume Open
Journal Subjects:
Engineering, Bioengineering and Biomedical Engineering, Biomedical Electronics, Life Sciences, Biophysics, Medicine, Biomedical Engineering, Physics, Spectroscopy and Metrology
Journal RSS Feed