Close

1. Identity statement
Reference TypeJournal Article
Sitemtc-m16.sid.inpe.br
Holder Codeisadg {BR SPINPE} ibi 8JMKD3MGPCW/3DT298S
Identifier6qtX3pFwXQZ3r59YDa/JsEno
Repositorysid.inpe.br/iris@1916/2005/12.14.14.51
Last Update2006:09.08.12.52.39 (UTC) administrator
Metadata Repositorysid.inpe.br/iris@1916/2005/12.14.14.51.03
Metadata Last Update2018:06.05.01.16.27 (UTC) administrator
Secondary KeyINPE-14155-PRE/9288
ISSN0102-7786
Citation KeyHarterCamp:2005:ReFeNe
TitleRecurrent and feedforward neural networks trained with cross validation scheme applied to the data assimilation in chaotic dynamics
Year2005
Secondary Date20060908
MonthDec.
Access Date2024, May 17
Secondary TypePRE PN
Number of Files1
Size913 KiB
2. Context
Author1 Harter, Fabrício Pereira
2 Campos Velho, Haroldo Fraga de
Resume Identifier1
2 8JMKD3MGP5W/3C9JHC3
Group1 LAC-INPE-MCT-BR
2 LAC-INPE-MCT-BR
Affiliation1 Instituto Nacional de Pesquisas Espaciais (INPE)
2 Instituto Nacional de Pesquisas Espaciais (INPE)
e-Mail Addressatus@cptec.inpe.br
JournalRevista Brasileira de Meteorologia
Volume20
Number3
Pages411-420
History (UTC)2006-09-20 13:36:02 :: simone -> administrator ::
2008-06-10 22:40:53 :: administrator -> banon ::
2010-05-14 14:20:50 :: banon -> administrator ::
2018-06-05 01:16:27 :: administrator -> marciana :: 2005
3. Content and structure
Is the master or a copy?is the master
Content Stagecompleted
Transferable1
Content TypeExternal Contribution
Keywordscross validation
data assimilation
Lorenz system
recurrent neural networks
AbstractArtificial Neural network (ANN) is a new approach for data assimilation process. The performance of two feedforward (multilayer perceptron and radial basis function), and two recurrent (Elman and Jordan) ANNs are analyzed. The Lorenzs system under chaotic regime is used as a test problem. These four ANNs were trained for emulating a Kalman filter using cross validation scheme. Multilayer Perceptron and Elman ANNs show better results among ANNs tested. The results obtained encouraging the application of the ANNs as na assimilation technique. Rede Neural Artificial é uma nova abordagem para assimilação de dados. Neste artigo é analisado o desempenho de duas redes feedforward (perceptron de múltiplas camadas e função de base radial) e duas redes recorrentes (Elman e Jordan). O sistema de Lorenz sob regime caótico é usado como um problema teste. As redes foram treinadas para emular um filtro de Kalman, usando a técnica de validação cruzada. O perceptron de múltiplas camadas e a rede de Elman apresentaram melhor desempenho entre as redes testadas. Os resultados encorajam a investigação de redes neurais como uma técnica para assimilação de dados em previsão de tempo operacional..
AreaMET
Arrangementurlib.net > BDMCI > Fonds > Produção anterior à 2021 > LABAC > Recurrent and feedforward...
doc Directory Contentaccess
source Directory Contentthere are no files
agreement Directory Contentthere are no files
4. Conditions of access and use
data URLhttp://urlib.net/ibi/6qtX3pFwXQZ3r59YDa/JsEno
zipped data URLhttp://urlib.net/zip/6qtX3pFwXQZ3r59YDa/JsEno
Languageen
Target FileHarter.Recurrent.pdf
User Groupadministrator
banon
simone
Visibilityshown
Copy HolderSID/SCD
Archiving Policyallowpublisher allowfinaldraft
Read Permissionallow from all
5. Allied materials
Next Higher Units8JMKD3MGPCW/3ESGTTP
Citing Item Listsid.inpe.br/mtc-m21/2012/07.13.14.49.40 3
DisseminationPORTALCAPES; SCIELO.
Host Collectionsid.inpe.br/banon/2003/08.15.17.40
6. Notes
Empty Fieldsalternatejournal archivist callnumber copyright creatorhistory descriptionlevel documentstage doi electronicmailaddress format isbn label lineage mark mirrorrepository nextedition notes orcid parameterlist parentrepositories previousedition previouslowerunit progress project readergroup rightsholder schedulinginformation secondarymark session shorttitle sponsor subject tertiarymark tertiarytype typeofwork url versiontype
7. Description control
e-Mail (login)marciana
update 


Close