Optimization of the weights and asymmetric activation function family of neural network for time series forecasting

Detalhes bibliográficos
Autor(a) principal: Gomes, Gecynalda S. da S.
Data de Publicação: 2013
Outros Autores: Ludermir, Teresa B.
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Repositório Institucional da UFBA
Texto Completo: http://repositorio.ufba.br/ri/handle/ri/15001
Resumo: Texto completo: acesso restrito. p. 6438–6446
id UFBA-2_f46f8b65226599fe8720a6acd917e68b
oai_identifier_str oai:repositorio.ufba.br:ri/15001
network_acronym_str UFBA-2
network_name_str Repositório Institucional da UFBA
repository_id_str 1932
spelling Gomes, Gecynalda S. da S.Ludermir, Teresa B.Gomes, Gecynalda S. da S.Ludermir, Teresa B.2014-05-20T13:06:42Z20130957-4174http://repositorio.ufba.br/ri/handle/ri/15001v. 40, n. 16Texto completo: acesso restrito. p. 6438–6446The use of neural network models for time series forecasting has been motivated by experimental results that indicate high capacity for function approximation with good accuracy. Generally, these models use activation functions with fixed parameters. However, it is known that the choice of activation function strongly influences the complexity and neural network performance and that a limited number of activation functions has been used in general. We describe the use of an asymmetric activation functions family with free parameter for neural networks. We prove that the activation functions family defined, satisfies the requirements of the universal approximation theorem We present a methodology for global optimization of the activation functions family with free parameter and the connections between the processing units of the neural network. The main idea is to optimize, simultaneously, the weights and activation function used in a Multilayer Perceptron (MLP), through an approach that combines the advantages of simulated annealing, tabu search and a local learning algorithm. We have chosen two local learning algorithms: the backpropagation with momentum (BPM) and Levenberg–Marquardt (LM). The overall purpose is to improve performance in time series forecasting.Submitted by Suelen Reis (suziy.ellen@gmail.com) on 2014-05-20T13:06:42Z No. of bitstreams: 1 1-s2.0-S0957417413003515-main.pdf: 468498 bytes, checksum: e5534a2216f8bf51c4331c8250a40c88 (MD5)Made available in DSpace on 2014-05-20T13:06:42Z (GMT). No. of bitstreams: 1 1-s2.0-S0957417413003515-main.pdf: 468498 bytes, checksum: e5534a2216f8bf51c4331c8250a40c88 (MD5) Previous issue date: 2013http://dx.doi.org/10.1016/j.eswa.2013.05.053reponame:Repositório Institucional da UFBAinstname:Universidade Federal da Bahia (UFBA)instacron:UFBANeural networksAsymmetric activation functionFree parameterSimulated annealingTabu searchBPM algorithmLM algorithmTime seriesOptimization of the weights and asymmetric activation function family of neural network for time series forecastingExpert Systems with Applicationsinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/article10000-01-01info:eu-repo/semantics/openAccessengORIGINAL1-s2.0-S0957417413003515-main.pdf1-s2.0-S0957417413003515-main.pdfapplication/pdf468498https://repositorio.ufba.br/bitstream/ri/15001/1/1-s2.0-S0957417413003515-main.pdfe5534a2216f8bf51c4331c8250a40c88MD51LICENSElicense.txtlicense.txttext/plain1345https://repositorio.ufba.br/bitstream/ri/15001/2/license.txt0d4b811ef71182510d2015daa7c8a900MD52TEXT1-s2.0-S0957417413003515-main.pdf.txt1-s2.0-S0957417413003515-main.pdf.txtExtracted texttext/plain58392https://repositorio.ufba.br/bitstream/ri/15001/3/1-s2.0-S0957417413003515-main.pdf.txteb9a110bd18d31c9165c575ed1505687MD53ri/150012022-07-05 14:03:13.375oai:repositorio.ufba.br:ri/15001VGVybW8gZGUgTGljZW4/YSwgbj9vIGV4Y2x1c2l2bywgcGFyYSBvIGRlcD9zaXRvIG5vIFJlcG9zaXQ/cmlvIEluc3RpdHVjaW9uYWwgZGEgVUZCQS4KCiBQZWxvIHByb2Nlc3NvIGRlIHN1Ym1pc3M/byBkZSBkb2N1bWVudG9zLCBvIGF1dG9yIG91IHNldSByZXByZXNlbnRhbnRlIGxlZ2FsLCBhbyBhY2VpdGFyIAplc3NlIHRlcm1vIGRlIGxpY2VuP2EsIGNvbmNlZGUgYW8gUmVwb3NpdD9yaW8gSW5zdGl0dWNpb25hbCBkYSBVbml2ZXJzaWRhZGUgRmVkZXJhbCBkYSBCYWhpYSAKbyBkaXJlaXRvIGRlIG1hbnRlciB1bWEgYz9waWEgZW0gc2V1IHJlcG9zaXQ/cmlvIGNvbSBhIGZpbmFsaWRhZGUsIHByaW1laXJhLCBkZSBwcmVzZXJ2YT8/by4gCkVzc2VzIHRlcm1vcywgbj9vIGV4Y2x1c2l2b3MsIG1hbnQ/bSBvcyBkaXJlaXRvcyBkZSBhdXRvci9jb3B5cmlnaHQsIG1hcyBlbnRlbmRlIG8gZG9jdW1lbnRvIApjb21vIHBhcnRlIGRvIGFjZXJ2byBpbnRlbGVjdHVhbCBkZXNzYSBVbml2ZXJzaWRhZGUuCgogUGFyYSBvcyBkb2N1bWVudG9zIHB1YmxpY2Fkb3MgY29tIHJlcGFzc2UgZGUgZGlyZWl0b3MgZGUgZGlzdHJpYnVpPz9vLCBlc3NlIHRlcm1vIGRlIGxpY2VuP2EgCmVudGVuZGUgcXVlOgoKIE1hbnRlbmRvIG9zIGRpcmVpdG9zIGF1dG9yYWlzLCByZXBhc3NhZG9zIGEgdGVyY2Vpcm9zLCBlbSBjYXNvIGRlIHB1YmxpY2E/P2VzLCBvIHJlcG9zaXQ/cmlvCnBvZGUgcmVzdHJpbmdpciBvIGFjZXNzbyBhbyB0ZXh0byBpbnRlZ3JhbCwgbWFzIGxpYmVyYSBhcyBpbmZvcm1hPz9lcyBzb2JyZSBvIGRvY3VtZW50bwooTWV0YWRhZG9zIGVzY3JpdGl2b3MpLgoKIERlc3RhIGZvcm1hLCBhdGVuZGVuZG8gYW9zIGFuc2Vpb3MgZGVzc2EgdW5pdmVyc2lkYWRlIGVtIG1hbnRlciBzdWEgcHJvZHU/P28gY2llbnQ/ZmljYSBjb20gCmFzIHJlc3RyaT8/ZXMgaW1wb3N0YXMgcGVsb3MgZWRpdG9yZXMgZGUgcGVyaT9kaWNvcy4KCiBQYXJhIGFzIHB1YmxpY2E/P2VzIHNlbSBpbmljaWF0aXZhcyBxdWUgc2VndWVtIGEgcG9sP3RpY2EgZGUgQWNlc3NvIEFiZXJ0bywgb3MgZGVwP3NpdG9zIApjb21wdWxzP3Jpb3MgbmVzc2UgcmVwb3NpdD9yaW8gbWFudD9tIG9zIGRpcmVpdG9zIGF1dG9yYWlzLCBtYXMgbWFudD9tIGFjZXNzbyBpcnJlc3RyaXRvIAphbyBtZXRhZGFkb3MgZSB0ZXh0byBjb21wbGV0by4gQXNzaW0sIGEgYWNlaXRhPz9vIGRlc3NlIHRlcm1vIG4/byBuZWNlc3NpdGEgZGUgY29uc2VudGltZW50bwogcG9yIHBhcnRlIGRlIGF1dG9yZXMvZGV0ZW50b3JlcyBkb3MgZGlyZWl0b3MsIHBvciBlc3RhcmVtIGVtIGluaWNpYXRpdmFzIGRlIGFjZXNzbyBhYmVydG8uCg==Repositório InstitucionalPUBhttp://192.188.11.11:8080/oai/requestopendoar:19322022-07-05T17:03:13Repositório Institucional da UFBA - Universidade Federal da Bahia (UFBA)false
dc.title.pt_BR.fl_str_mv Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
dc.title.alternative.pt_BR.fl_str_mv Expert Systems with Applications
title Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
spellingShingle Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
Gomes, Gecynalda S. da S.
Neural networks
Asymmetric activation function
Free parameter
Simulated annealing
Tabu search
BPM algorithm
LM algorithm
Time series
title_short Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
title_full Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
title_fullStr Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
title_full_unstemmed Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
title_sort Optimization of the weights and asymmetric activation function family of neural network for time series forecasting
author Gomes, Gecynalda S. da S.
author_facet Gomes, Gecynalda S. da S.
Ludermir, Teresa B.
author_role author
author2 Ludermir, Teresa B.
author2_role author
dc.contributor.author.fl_str_mv Gomes, Gecynalda S. da S.
Ludermir, Teresa B.
Gomes, Gecynalda S. da S.
Ludermir, Teresa B.
dc.subject.por.fl_str_mv Neural networks
Asymmetric activation function
Free parameter
Simulated annealing
Tabu search
BPM algorithm
LM algorithm
Time series
topic Neural networks
Asymmetric activation function
Free parameter
Simulated annealing
Tabu search
BPM algorithm
LM algorithm
Time series
description Texto completo: acesso restrito. p. 6438–6446
publishDate 2013
dc.date.issued.fl_str_mv 2013
dc.date.accessioned.fl_str_mv 2014-05-20T13:06:42Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://repositorio.ufba.br/ri/handle/ri/15001
dc.identifier.issn.none.fl_str_mv 0957-4174
dc.identifier.number.pt_BR.fl_str_mv v. 40, n. 16
identifier_str_mv 0957-4174
v. 40, n. 16
url http://repositorio.ufba.br/ri/handle/ri/15001
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.source.pt_BR.fl_str_mv http://dx.doi.org/10.1016/j.eswa.2013.05.053
dc.source.none.fl_str_mv reponame:Repositório Institucional da UFBA
instname:Universidade Federal da Bahia (UFBA)
instacron:UFBA
instname_str Universidade Federal da Bahia (UFBA)
instacron_str UFBA
institution UFBA
reponame_str Repositório Institucional da UFBA
collection Repositório Institucional da UFBA
bitstream.url.fl_str_mv https://repositorio.ufba.br/bitstream/ri/15001/1/1-s2.0-S0957417413003515-main.pdf
https://repositorio.ufba.br/bitstream/ri/15001/2/license.txt
https://repositorio.ufba.br/bitstream/ri/15001/3/1-s2.0-S0957417413003515-main.pdf.txt
bitstream.checksum.fl_str_mv e5534a2216f8bf51c4331c8250a40c88
0d4b811ef71182510d2015daa7c8a900
eb9a110bd18d31c9165c575ed1505687
bitstream.checksumAlgorithm.fl_str_mv MD5
MD5
MD5
repository.name.fl_str_mv Repositório Institucional da UFBA - Universidade Federal da Bahia (UFBA)
repository.mail.fl_str_mv
_version_ 1808459478697246720

Registros relacionados