Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor
In the process of online prediction of multivariable non-stationary time series by kernel extreme learning machine (KELM), the dynamic characteristics of the system which are difficult to determine have always posed a big problem. We propose an online sequential prediction model with an adaptive for...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9204612/ |
id |
doaj-bdf4062dc18748e3875e58c9344b45b8 |
---|---|
record_format |
Article |
spelling |
doaj-bdf4062dc18748e3875e58c9344b45b82021-03-30T03:56:34ZengIEEEIEEE Access2169-35362020-01-01817595817597110.1109/ACCESS.2020.30260099204612Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting FactorJinling Dai0https://orcid.org/0000-0001-9931-6707Aiqiang Xu1https://orcid.org/0000-0001-9109-3385Xing Liu2Chao Yu3Yangyong Wu4Naval Aviation University, Yantai, ChinaNaval Aviation University, Yantai, ChinaNaval Aviation University, Yantai, China92313 Troops, Jiyuan, ChinaNaval Aviation University, Yantai, ChinaIn the process of online prediction of multivariable non-stationary time series by kernel extreme learning machine (KELM), the dynamic characteristics of the system which are difficult to determine have always posed a big problem. We propose an online sequential prediction model with an adaptive forgetting factor (AFF) for multivariable time series to solve this problem. The multivariable time series instead of variable itself is reconstructed firstly. AFF is introduced into the objective function and can be adjusted iteratively and adaptively with the system changes. As a result, higher weight can be allocated for the fresh and more important samples while the old failure samples can be quickly forgotten. The model sparsification uses a fast leave-one-out cross-validation (FLOO-CV) method to set a prediction error threshold so that samples can be selected conditionally to form a dictionary. Besides, the dictionary parameters, including AFF and kernel parameters, are recursively updated simultaneously without increasing calculation complexity. The experimental results show that, compared with four fashionable KLEM methods, the proposed AFF-OSKELM has a better dynamic tracking ability and adaptability. Moreover, compared with single variable prediction, the spatial reconstructed multivariable has higher prediction accuracy and stability.https://ieeexplore.ieee.org/document/9204612/Kernel extreme learning machineadaptive forgetting factorfast leave-one-out cross-validationonline predictionmultivariable time series |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Jinling Dai Aiqiang Xu Xing Liu Chao Yu Yangyong Wu |
spellingShingle |
Jinling Dai Aiqiang Xu Xing Liu Chao Yu Yangyong Wu Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor IEEE Access Kernel extreme learning machine adaptive forgetting factor fast leave-one-out cross-validation online prediction multivariable time series |
author_facet |
Jinling Dai Aiqiang Xu Xing Liu Chao Yu Yangyong Wu |
author_sort |
Jinling Dai |
title |
Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor |
title_short |
Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor |
title_full |
Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor |
title_fullStr |
Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor |
title_full_unstemmed |
Online Sequential Model for Multivariate Time Series Prediction With Adaptive Forgetting Factor |
title_sort |
online sequential model for multivariate time series prediction with adaptive forgetting factor |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
In the process of online prediction of multivariable non-stationary time series by kernel extreme learning machine (KELM), the dynamic characteristics of the system which are difficult to determine have always posed a big problem. We propose an online sequential prediction model with an adaptive forgetting factor (AFF) for multivariable time series to solve this problem. The multivariable time series instead of variable itself is reconstructed firstly. AFF is introduced into the objective function and can be adjusted iteratively and adaptively with the system changes. As a result, higher weight can be allocated for the fresh and more important samples while the old failure samples can be quickly forgotten. The model sparsification uses a fast leave-one-out cross-validation (FLOO-CV) method to set a prediction error threshold so that samples can be selected conditionally to form a dictionary. Besides, the dictionary parameters, including AFF and kernel parameters, are recursively updated simultaneously without increasing calculation complexity. The experimental results show that, compared with four fashionable KLEM methods, the proposed AFF-OSKELM has a better dynamic tracking ability and adaptability. Moreover, compared with single variable prediction, the spatial reconstructed multivariable has higher prediction accuracy and stability. |
topic |
Kernel extreme learning machine adaptive forgetting factor fast leave-one-out cross-validation online prediction multivariable time series |
url |
https://ieeexplore.ieee.org/document/9204612/ |
work_keys_str_mv |
AT jinlingdai onlinesequentialmodelformultivariatetimeseriespredictionwithadaptiveforgettingfactor AT aiqiangxu onlinesequentialmodelformultivariatetimeseriespredictionwithadaptiveforgettingfactor AT xingliu onlinesequentialmodelformultivariatetimeseriespredictionwithadaptiveforgettingfactor AT chaoyu onlinesequentialmodelformultivariatetimeseriespredictionwithadaptiveforgettingfactor AT yangyongwu onlinesequentialmodelformultivariatetimeseriespredictionwithadaptiveforgettingfactor |
_version_ |
1724182663176126464 |