Cover Image

Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data

Purpose Deep Neural Networks (DNNs) typically require enormous labeled training samples to achieve optimum performance. Therefore, numerous forms of data augmentation techniques are employed to compensate for the lack of training samples. Methods In this paper, a data augmentation technique named...

Full description

Saved in:
Bibliographic Details
Main Authors: Atik, Faysal, W. K., Ngui, M. H., Lim, M. S., Leong
Format: Article
Language:English
Published: Springer Link 2022
Subjects:
TJ Mechanical engineering and machinery
Online Access:http://umpir.ump.edu.my/id/eprint/35522/1/Ensemble%20Augmentation%20for.pdf
http://umpir.ump.edu.my/id/eprint/35522/
https://doi.org/10.1007/s42417-022-00683-w
https://doi.org/10.1007/s42417-022-00683-w
Tags: Add Tag
No Tags, Be the first to tag this record!
id my.ump.umpir.35522
record_format eprints
spelling my.ump.umpir.355222022-11-03T02:11:29Z http://umpir.ump.edu.my/id/eprint/35522/ Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data Atik, Faysal W. K., Ngui M. H., Lim M. S., Leong TJ Mechanical engineering and machinery Purpose Deep Neural Networks (DNNs) typically require enormous labeled training samples to achieve optimum performance. Therefore, numerous forms of data augmentation techniques are employed to compensate for the lack of training samples. Methods In this paper, a data augmentation technique named ensemble augmentation is proposed to generate real-like samples. This augmentation method uses the power of white noise added in ensembles to the original samples to generate real-like samples. After averaging the signal with ensembles, a new signal is obtained that contains the characteristics of the original signal. The parameters for the ensemble augmentation are validated using a simulated signal. The proposed method is evaluated by 10 class-bearing vibration data using three Transfer Learning (TL) models, namely, Inception-V3, MobileNet-V2, and ResNet50. The outputs from the proposed method are compared with no augmentation and different augmentation techniques. Results The results showed that the classifiers with the ensemble augmentation have higher validation and test accuracy than all the other augmentation techniques. The robustness assessment conducted with noisy test samples and test samples from different loads showed that the classifiers could obtain much higher robustness when trained with samples from ensemble augmentation. Conclusion The proposed data augmentation technique can be applied to 1-D time series data to achieve robust classifiers. Springer Link 2022 Article PeerReviewed pdf en http://umpir.ump.edu.my/id/eprint/35522/1/Ensemble%20Augmentation%20for.pdf Atik, Faysal and W. K., Ngui and M. H., Lim and M. S., Leong (2022) Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data. Journal of Vibration Engineering & Technologies. pp. 1-25. ISSN 2523-3939 https://doi.org/10.1007/s42417-022-00683-w https://doi.org/10.1007/s42417-022-00683-w
institution Universiti Malaysia Pahang
building UMP Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaysia Pahang
content_source UMP Institutional Repository
url_provider http://umpir.ump.edu.my/
language English
topic TJ Mechanical engineering and machinery
spellingShingle TJ Mechanical engineering and machinery
Atik, Faysal
W. K., Ngui
M. H., Lim
M. S., Leong
Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
description Purpose Deep Neural Networks (DNNs) typically require enormous labeled training samples to achieve optimum performance. Therefore, numerous forms of data augmentation techniques are employed to compensate for the lack of training samples. Methods In this paper, a data augmentation technique named ensemble augmentation is proposed to generate real-like samples. This augmentation method uses the power of white noise added in ensembles to the original samples to generate real-like samples. After averaging the signal with ensembles, a new signal is obtained that contains the characteristics of the original signal. The parameters for the ensemble augmentation are validated using a simulated signal. The proposed method is evaluated by 10 class-bearing vibration data using three Transfer Learning (TL) models, namely, Inception-V3, MobileNet-V2, and ResNet50. The outputs from the proposed method are compared with no augmentation and different augmentation techniques. Results The results showed that the classifiers with the ensemble augmentation have higher validation and test accuracy than all the other augmentation techniques. The robustness assessment conducted with noisy test samples and test samples from different loads showed that the classifiers could obtain much higher robustness when trained with samples from ensemble augmentation. Conclusion The proposed data augmentation technique can be applied to 1-D time series data to achieve robust classifiers.
format Article
author Atik, Faysal
W. K., Ngui
M. H., Lim
M. S., Leong
author_facet Atik, Faysal
W. K., Ngui
M. H., Lim
M. S., Leong
author_sort Atik, Faysal
title Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
title_short Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
title_full Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
title_fullStr Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
title_full_unstemmed Ensemble Augmentation for Deep Neural Networks Using 1-D Time Series Vibration Data
title_sort ensemble augmentation for deep neural networks using 1-d time series vibration data
publisher Springer Link
publishDate 2022
url http://umpir.ump.edu.my/id/eprint/35522/1/Ensemble%20Augmentation%20for.pdf
http://umpir.ump.edu.my/id/eprint/35522/
https://doi.org/10.1007/s42417-022-00683-w
https://doi.org/10.1007/s42417-022-00683-w
_version_ 1748703317815459840
score 13.211869

Similar Items