View in EDS

Optimizing echo state network with backtracking search optimization algorithm for time series forecasting.

Saved in:
Bibliographic Details
Title: Optimizing echo state network with backtracking search optimization algorithm for time series forecasting.
Authors: Wang, Zhigang1,2 (AUTHOR) wangzhigang2@ygsoft.com, Zeng, Yu-Rong3 (AUTHOR) zyr@hbue.edu.cn, Wang, Sirui1 (AUTHOR) ruisiwang@foxmail.com, Wang, Lin1 (AUTHOR) wanglin@hust.edu.cn
Source: Engineering Applications of Artificial Intelligence. May2019, Vol. 81, p117-132. 16p.
Subjects: Mathematical optimization, Search algorithms, Evolutionary algorithms, Regression analysis, Echo
Abstract: The echo state network (ESN) is a state-of-the art reservoir computing approach, which is particularly effective for time series forecasting problems because it is coupled with a time parameter. However, the linear regression algorithm commonly used to compute the output weights of ESN could usually cause the trained network over-fitted and thus obtain unsatisfactory results. To overcome the problem, we present four optimized ESNs that are based on the backtracking search optimization algorithm (BSA) or its variants to improve generalizability. Concretely, we utilize BSA and its variants to determine the most appropriate output weights of ESN given that the optimization problem is complex while BSA is a novel evolutionary algorithm that effectively unscrambles optimal solutions in complex spaces. The three BSA variants, namely, adaptive population selection scheme (APSS)–BSA, adaptive mutation factor strategy (AMFS)–BSA, and APSS&AMFS–BSA, were designed to further improve the performance of BSA. Time series forecasting experiments were performed using two real-life time series. The experimental results of the optimized ESNs were compared with those of the basic ESN without optimization, and the two other comparison approaches, as well as the other existing approaches. Experimental results showed that (a) the results of the optimized ESNs are more accurate than that of basic ESN and (b) APSS&AMFS–BSA–ESN nearly outperforms basic ESN, the three other optimized ESNs, the two comparison approaches, and other existing optimization approaches. • Optimized echo state networks are proposed for effective time series forecasting. • Three backtracking search algorithm variants are proposed for better performance. • Backtracking search algorithms help avoiding overfitting of echo state networks. • The proposed approach named APSS&AMFS–BSA-ESN is the best for two cases. [ABSTRACT FROM AUTHOR]
Copyright of Engineering Applications of Artificial Intelligence is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 136730600
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Optimizing echo state network with backtracking search optimization algorithm for time series forecasting.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Wang%2C+Zhigang%22">Wang, Zhigang</searchLink><relatesTo>1,2</relatesTo> (AUTHOR)<i> wangzhigang2@ygsoft.com</i><br /><searchLink fieldCode="AR" term="%22Zeng%2C+Yu-Rong%22">Zeng, Yu-Rong</searchLink><relatesTo>3</relatesTo> (AUTHOR)<i> zyr@hbue.edu.cn</i><br /><searchLink fieldCode="AR" term="%22Wang%2C+Sirui%22">Wang, Sirui</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> ruisiwang@foxmail.com</i><br /><searchLink fieldCode="AR" term="%22Wang%2C+Lin%22">Wang, Lin</searchLink><relatesTo>1</relatesTo> (AUTHOR)<i> wanglin@hust.edu.cn</i>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22Engineering+Applications+of+Artificial+Intelligence%22">Engineering Applications of Artificial Intelligence</searchLink>. May2019, Vol. 81, p117-132. 16p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Mathematical+optimization%22">Mathematical optimization</searchLink><br /><searchLink fieldCode="DE" term="%22Search+algorithms%22">Search algorithms</searchLink><br /><searchLink fieldCode="DE" term="%22Evolutionary+algorithms%22">Evolutionary algorithms</searchLink><br /><searchLink fieldCode="DE" term="%22Regression+analysis%22">Regression analysis</searchLink><br /><searchLink fieldCode="DE" term="%22Echo%22">Echo</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: The echo state network (ESN) is a state-of-the art reservoir computing approach, which is particularly effective for time series forecasting problems because it is coupled with a time parameter. However, the linear regression algorithm commonly used to compute the output weights of ESN could usually cause the trained network over-fitted and thus obtain unsatisfactory results. To overcome the problem, we present four optimized ESNs that are based on the backtracking search optimization algorithm (BSA) or its variants to improve generalizability. Concretely, we utilize BSA and its variants to determine the most appropriate output weights of ESN given that the optimization problem is complex while BSA is a novel evolutionary algorithm that effectively unscrambles optimal solutions in complex spaces. The three BSA variants, namely, adaptive population selection scheme (APSS)–BSA, adaptive mutation factor strategy (AMFS)–BSA, and APSS&AMFS–BSA, were designed to further improve the performance of BSA. Time series forecasting experiments were performed using two real-life time series. The experimental results of the optimized ESNs were compared with those of the basic ESN without optimization, and the two other comparison approaches, as well as the other existing approaches. Experimental results showed that (a) the results of the optimized ESNs are more accurate than that of basic ESN and (b) APSS&AMFS–BSA–ESN nearly outperforms basic ESN, the three other optimized ESNs, the two comparison approaches, and other existing optimization approaches. • Optimized echo state networks are proposed for effective time series forecasting. • Three backtracking search algorithm variants are proposed for better performance. • Backtracking search algorithms help avoiding overfitting of echo state networks. • The proposed approach named APSS&AMFS–BSA-ESN is the best for two cases. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of Engineering Applications of Artificial Intelligence is the property of Pergamon Press - An Imprint of Elsevier Science and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=136730600
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1016/j.engappai.2019.02.009
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 16
        StartPage: 117
    Subjects:
      – SubjectFull: Mathematical optimization
        Type: general
      – SubjectFull: Search algorithms
        Type: general
      – SubjectFull: Evolutionary algorithms
        Type: general
      – SubjectFull: Regression analysis
        Type: general
      – SubjectFull: Echo
        Type: general
    Titles:
      – TitleFull: Optimizing echo state network with backtracking search optimization algorithm for time series forecasting.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Wang, Zhigang
      – PersonEntity:
          Name:
            NameFull: Zeng, Yu-Rong
      – PersonEntity:
          Name:
            NameFull: Wang, Sirui
      – PersonEntity:
          Name:
            NameFull: Wang, Lin
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 05
              Text: May2019
              Type: published
              Y: 2019
          Identifiers:
            – Type: issn-print
              Value: 09521976
          Numbering:
            – Type: volume
              Value: 81
          Titles:
            – TitleFull: Engineering Applications of Artificial Intelligence
              Type: main
ResultId 1