Analysis and Improvement Approach of the Impact of Data Disturbance on LSTM Prediction Algorithm

Authors

  • Xin Yidan School of Automation and Electrical Engineering, Xi’an University of Technology, Xi’an, China
  • Hu Shaolin Automation School, Guangdong University of Petrochemical Technology, Maoming, China
  • Yang Guotao School of Automation and Electrical Engineering, Xi’an University of Technology, Xi’an, China

DOI:

https://doi.org/10.14738/tecs.115.15411

Keywords:

LSTM, Outlier, Fault-tolerant filtering algorithm, Fault tolerance

Abstract

If the sampling data with noise or outliers are used to train the long-short-term memory (short as LSTM) network, whether the perturbations and outliers in sampling data affect the training performance and prediction accuracy of LSTM networks is a key problem. This paper analyzed the impact on the LSTM prediction effect when using perturbations and isolated/patchy outliers involved data for network training and prediction. The results showed that the prediction accuracy decreases as the amplitude of the perturbations and the range of outliers increase. In order to overcome this effect, an improved method of Pre-set Outliers Tolerant Filter is proposed, and an Outliers-Tolerant Multi-LSTM model, in short, the OTML model, is established. The prediction effect of the proposed model is compared with that of the LSTM model without the filter and that of the LSTM model with a mean filter. Comparison results showed that the OTML prediction model proposed in this paper can eliminate the influence of random noise and isolated outliers thoroughly. When meeting the patchy outliers whose width is smaller than the radius of the window, the OTML prediction model can filter them out too, so as to realize the high-fidelity prediction of the data.

References

Karimanzira, D., Ritzau, L., & Emde, K. (2022). Catchment Area Multi-Streamflow Multiple Hours Ahead Forecast Based on Deep Learning. Transactions on Engineering and Computing Sciences, 10(5), 15–29. https://doi.org/10.14738/tmlai.105.13049

namaswa, S., githiri, john., Mariita, N.., k’Orowe, M..., & Njiru, N. (2021). Application of Data-Driven Discovery Machine Learning Algorithms in Predicting Geothermal Reservoir Temperature from Geophysical Resistivity Method. Transactions on Engineering and Computing Sciences, 9(4), 1–12. https://doi.org/10.14738/tmlai.94.10281

Zhang Lewen. Research on key technology of multi-component gas monitoring in heating furnace combustion process[D]. University of Science and Technology of China, 2022.DOI: 10.27517/d.cnki.gzkju.2022.000425.)

WANG Xin, WU Ji, LIU Chao, YANG Haiyan, DU Yanli, NIU Wensheng. Failure time series prediction based on LSTM recurrent neural network[J]. Journal of Beihang University,2018,44(04): 772-784.DOI:10.13700/j.bh.1001-5965.2017.0285.)

WANG Lipeng, LIANG Qi, ZOU Xianli, et al. Spectral noise reduction analysis method based on optimized moving average filtering[J]. Electrical Engineering,2018,19(11):43-45+54.)

YUE Haiyan, YAO Hong, ZHOU Zunsi. Comparison and simulation of filtering algorithms[J]. Wireless Internet Technology,2017(08):103-104.)

HU Shunbo. Fusion interpolation algorithm of B-spline filtering and GPVE in image registration[J]. Computer Applications,2011,31(08):2225-2228.)

J. Boudet, E. Lévêque, P. Borgnat, A. Cahuzac, M.C. Jacob. A Kalman filter adapted to the estimation of mean gradients in the large-eddy simulation of unsteady turbulent flows. Computers & Fluids, Volume 127, 2016, Pages 65-77, ISSN 0045-7930, https://doi.org/10.1016/j.compfluid.2015.12.006

CHEN Long. Prediction method of process variables driven by complex industrial data[D]. Dalian University of Technology, 2019.DOI: 10.26991/d.cnki.gdllu.2019.004119.)

Sheng Chunyang. Prediction method and application of secondary energy system of steel enterprises based on data[D]. Dalian University of Technology,2014.)

Wang Zhilan, Zhao Hongjie, Zhao Fan, et al. Establishment of remaining life prediction model of an inertial guide system based on convolutional neural network and filter fusion algorithm[J]. Aerospace Defense,2023,6(01):70-77.)

RONG Jing, LIU Lilong, WEN Hongyan, et al. Application of GM (1,1)-AR prediction model based on Kalman filter in deformation prediction[J]. Journal of Guilin University of Technology,2018,38(02):301-305.)

YI Lirong, WANG Shaoyu, YIN Lili, YANG Qing, GU Xin. Prediction of industrial sensor time series data based on multivariate LSTM[J]. Intelligent Computer and Application,2018,8(05):13-16.)

ZHU Feilin, HOU Tiantian, LIANG Yifan, REN Ruijie. Long-term runoff prediction model in reservoir based on VMD-LSTM[J/OL]. Hydroelectric Power:1-8[2023-05-25]. http://kns.cnki.net/kcms/detail/11.1845.TV.20230423.1719.004.html

Downloads

Published

2023-09-10

How to Cite

Yidan, X., Shaolin, H., & Guotao, Y. (2023). Analysis and Improvement Approach of the Impact of Data Disturbance on LSTM Prediction Algorithm. Transactions on Engineering and Computing Sciences, 11(5), 1–15. https://doi.org/10.14738/tecs.115.15411