Deep Echo State Network Pruning Algorithm Based on Detrended Multiple Cross-correlation

NAVIGATE

Table of Contents

STATISTICS

Viewed230

Downloads195

Deep Echo State Network Pruning Algorithm Based on Detrended Multiple Cross-correlation

[HTML] PDF下载 (195)

[1]SUN Xiaochuan,WANG Yu,LI Yingqi,et al.Deep Echo State Network Pruning Algorithm Based on Detrended Multiple Cross-correlation[J].Journal of Zhengzhou University (Engineering Science),2024,45(04):38-45.[doi:10.13705/ j.issn.1671-6833.2024.04.005]

Copy

Journal of Zhengzhou University (Engineering Science)[ISSN 1671-6833/CN 41-1339/T] Volume: 45 Number of periods: 2024 04 Page number: 38-45 Column: Public date: 2024-06-16

Title:: Deep Echo State Network Pruning Algorithm Based on Detrended Multiple Cross-correlation

Author(s):: SUN Xiaochuan¹; 2; WANG Yu¹; 2; LI Yingqi¹; 2; HUANG Tianyu¹; 2; 1.College of Artificial Intelligence, North China University of Science and Technology, Tangshan 063210, China; 2.Hebei Provincial Key Laboratory of Industrial Intelligent Perception, Tangshan 063210, China

Keywords:: deep echo state network; structure optimization; pruning; detrended multiple cross-correlation; time series prediction

CLC:: TP183

DOI:: 10.13705/ j.issn.1671-6833.2024.04.005

Abstract:: To address the problem of undesirable prediction accuracy of a deep echo state network caused by redun dant structures in the reservoirs, a pruning algorithm for the deep echo state network based on detrended multiple cross-correlation was proposed. Firstly, according to the detrended covariance function and the detrended variance function, the detrended cross-correlation coefficient between each two neurons in the selected reservoirs in turn was calculated, and the detrended cross-correlation matrix was constructed. Based on this matrix, the detrended multi ple cross-correlation between a selected neuron and all remaining neurons in this reservoir could be evaluated. Sub sequently, the connections from the highly correlated neurons in each reservoir to the output layer were pruned se quentially, thus removing redundant components in the network. Finally, the network after pruning was retrained by least squares regression to obtain the optimal deep echo state network topology. Simulation results showed that the prediction accuracy and memory capacity of the deep echo state network optimized by the proposed algorithm on Mackey-Glass time series were improved by 89.80% and 30.93%, respectively, and on Call time series by 14.34% and 0.10%, respectively.

References:: [1] BARREDO ARRIETA A, GIL-LOPEZ S, LAÑA I, et al. On the post-hoc explainability of deep echo state net works for time series forecasting, image and video classi fication[J]. Neural Computing and Applications, 2022, 34(13): 10257-10277.
[2] 李北明, 金荣璐, 徐召飞, 等. 基于特征蒸馏的改进 Ghost-YOLOv5红外目标检测算法[J]. 郑州大学学报 (工学版), 2022, 43(1): 20-26.

LI B M, JIN R L, XU Z F, et al. An improved Ghost YOLOv5 infrared target detection algorithm based on fea ture distillation[J]. Journal of Zhengzhou University (Engineering Science), 2022, 43(1): 20-26.

[3] SHEN Q Y, ZHANG H W, MAO Y. Improving deep echo state network with neuronal similarity-based iterative pruning merging algorithm[J]. Applied Sciences, 2023, 13(5): 2918.

[4] GAO R B, LI R L, HU M H, et al. Dynamic ensemble deep echo state network for significant wave height forecasting[J]. Applied Energy, 2023, 329: 120261.

[5] ZEBENDE G F, DA SILVA FILHO A M. Detrended multiple cross-correlation coefficient[J]. Physica A: Sta tistical Mechanics and its Applications, 2018, 510: 91-97.

[6] WANG F, XU J, FAN Q J. Statistical properties of the detrended multiple cross-correlation coefficient[J]. Com munications in Nonlinear Science and Numerical Simula tion, 2021, 99: 105781.

[7] DE ALMEIDA BRITO A, DE ARAÚJO H A, ZEBENDE G F. Detrended multiple cross-correlation coefficient ap plied to solar radiation, air temperature and relative hu midity[J]. Scientific Reports, 2019, 9: 19764.

[8] GALLICCHIO C, MICHELI A. Echo state property of deep reservoir computing networks[J]. Cognitive Compu tation, 2017, 9(3): 337-350.

[9] BEN-SALHA O, MOKNI K. Detrended cross-correlation analysis in quantiles between oil price and the US stock market[J]. Energy, 2022, 242: 122918.

[10] BARLACCHI G, DE NADAI M, LARCHER R, et al. A multi-source dataset of urban life in the city of Milan and the Province of Trentino[J]. Scientific Data, 2015, 2: 150055.

[11] LI D Y, LIU F, QIAO J F, et al. Structure optimization for echo state network based on contribution[J]. Tsing hua Science and Technology, 2019, 24(1): 97-105.

[12] LI D, JIANG F X, CHEN M, et al. Multi-step-ahead wind speed forecasting based on a hybrid decomposition method and temporal convolutional networks[J]. Energy, 2022, 238: 121981.

[13]高金峰, 庞昊, 杜耀恒. 基于GRU网络的配电网故障数量等级预测方法[J]. 郑州大学学报(工学版), 2019, 40(5): 39-44.

GAO J F, PANG H, DU Y H. A method for predicting the number of faults in distribution network based on GRU neural network[J]. Journal of Zhengzhou University (Engineering Science), 2019, 40(5): 39-44.

[14] CHEN Z L, YANG C L, QIAO J F. The optimal design and application of LSTM neural network based on the hy brid coding PSO algorithm[J]. The Journal of Supercom puting, 2022, 78(5): 7227-7259.

[15] CUI M S. District heating load prediction algorithm based on bidirectional long short-term memory network model [J]. Energy, 2022, 254: 124283.

[16] HUANG Y P, YEN M F. A new perspective of perform ance comparison among machine learning algorithms for financial distress prediction[J]. Applied Soft Computing, 2019, 83: 105663.

[17] LI Y M, PENG T, HUA L, et al. Research and applica tion of an evolutionary deep learning model based on im proved grey wolf optimization algorithm and DBN-ELM for AQI prediction[J]. Sustainable Cities and Society, 2022, 87: 104209.

Similar References:

Memo

Last Update: 2024-06-14