Robust adaptive rescaled lncosh neural network regression toward time-series forecasting
Journal article
Yang, Yang, Zhou, Hu, Wu, Jinran, Ding, Zhe, Tian, Yu-Chu, Yue, Dong and Wang, You-Gan. (2023). Robust adaptive rescaled lncosh neural network regression toward time-series forecasting. IEEE Transactions on Systems, Man and Cybernetics: Systems. 53(9), pp. 5658-5669. https://doi.org/10.1109/TSMC.2023.3272880
Authors | Yang, Yang, Zhou, Hu, Wu, Jinran, Ding, Zhe, Tian, Yu-Chu, Yue, Dong and Wang, You-Gan |
---|---|
Abstract | In time series forecasting with outliers and random noise, parameter estimation in a neural network via minimizing the l2 loss is unreliable. Therefore, an adaptive rescaled lncosh loss function is proposed in this article to handle time series modeling with outliers and random noise. It overcomes the limitation of the single distribution of traditional loss functions and can switch among l1 , l2 , and the Huber losses. A tuning parameter in the loss function is estimated by using a “working” likelihood approach according to estimated residuals. From the proposed loss function, a robust adaptive rescaled lncosh neural network (RARLNN) regression model is developed for highly accurate predictions. In the training phase of the model, an iterative learning procedure is presented to estimate the tuning parameter and train the neural network in iterations. A new prediction interval construction method is also developed based on quantile theory. The proposed RARLNN model is applied to two groups of wind speed forecasting tasks. The results show that the proposed RARLNN model is more conducive to enhancing forecasting accuracy and stability from the perspectives of noise distribution and outliers. |
Keywords | outliers; prediction interval (PI); robust loss function; time series forecasting (TSF) |
Year | 2023 |
Journal | IEEE Transactions on Systems, Man and Cybernetics: Systems |
Journal citation | 53 (9), pp. 5658-5669 |
Publisher | Institute of Electrical and Electronics Engineers Inc. |
ISSN | 2168-2232 |
Digital Object Identifier (DOI) | https://doi.org/10.1109/TSMC.2023.3272880 |
Scopus EID | 2-s2.0-85161011517 |
Web address (URL) | https://ieeexplore.ieee.org/document/10130782 |
Open access | Published as non-open access |
Research or scholarly | Research |
Page range | 5658-5669 |
Publisher's version | License All rights reserved File Access Level Controlled |
Output status | Published |
Publication dates | |
Online | 23 May 2023 |
Publication process dates | |
Accepted | 30 Apr 2023 |
Deposited | 09 Aug 2023 |
https://acuresearchbank.acu.edu.au/item/8z7vx/robust-adaptive-rescaled-lncosh-neural-network-regression-toward-time-series-forecasting
Restricted files
Publisher's version
66
total views0
total downloads1
views this month0
downloads this month