Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures

Autores
Torra, Joel; Royo Esnal, Aritz; Chantre Balacca, Guillermo Ruben
Año de publicación
2016
Idioma
inglés
Tipo de recurso
documento de conferencia
Estado
versión publicada
Descripción
Artificial Neural Networks (ANNs) are machines with complexfunctional relations learnable with a limited amount of trainingdata emulating data processing functions of the brain. ANNs have ahigh potential applicability in ecological systems due to their capacity to describehighly non-linear relationships among variables. In this sense, they represent promising computational tools to accurately model weed emergence and therefore, for prediction purposes to improve weed control. But they have not beenwidely used with this aim, and so far, they have been proven to be useful only for one weed species (Avena fatua). The objectives of the present work were to develop an ANN model for ripgut brome (Bromus diandrus) for emergence prediction and to compare their predictive capability against already developed non-linear regression (NLR) models. Thermal-time and hydro-time were used as independent input variablesfor developing bivariate models. The accumulated proportion of seedling emergence was the output variable. A total of 1610 input/output data pairs corresponding to three years of data collection in two different field trials wereused in this study. A total of 16 different scenarios or emergence data sets (differing in sowing dates and soil management) were modeledto compare the goodness of fit (RMSE) by the two approaches. The ANN developed had three layers: one input layer, one hidden layer with 2 neurons, and one output layer. Both procedures, ANN and NLR, were able to predict satisfactorily B. diandrus emergence patterns. However, the ANN improved the fitting accuracy in 11 of the 16 scenarios with RMSE estimates 46% lower compared to NLR models. These results confirm that ANNs are powerful tools for modeling weed emergence, thus they could help improve IWM decision support systems.
Fil: Torra, Joel. Universidad de Lleida; España
Fil: Royo Esnal, Aritz. Universidad de Lleida; España
Fil: Chantre Balacca, Guillermo Ruben. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca. Centro de Recursos Naturales Renovables de la Zona Semiárida. Universidad Nacional del Sur. Centro de Recursos Naturales Renovables de la Zona Semiárida; Argentina. Universidad Nacional del Sur. Departamento de Agronomía; Argentina
7th International Weed Science Congress
Praga
República Checa
International Weed Science Society
Materia
THERMAL TIME
HIDROTIME
GERMINATION
SIGMOIDAL REGRESSION
Nivel de accesibilidad
acceso abierto
Condiciones de uso
https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
Repositorio
CONICET Digital (CONICET)
Institución
Consejo Nacional de Investigaciones Científicas y Técnicas
OAI Identificador
oai:ri.conicet.gov.ar:11336/264213

id CONICETDig_db61ce74f79a58128d71c9d3b02a9718
oai_identifier_str oai:ri.conicet.gov.ar:11336/264213
network_acronym_str CONICETDig
repository_id_str 3498
network_name_str CONICET Digital (CONICET)
spelling Modeling weed emergence: artifi cial neural networks versus non-linear regression proceduresTorra, JoelRoyo Esnal, AritzChantre Balacca, Guillermo RubenTHERMAL TIMEHIDROTIMEGERMINATIONSIGMOIDAL REGRESSIONhttps://purl.org/becyt/ford/4.1https://purl.org/becyt/ford/4https://purl.org/becyt/ford/1.2https://purl.org/becyt/ford/1Artificial Neural Networks (ANNs) are machines with complexfunctional relations learnable with a limited amount of trainingdata emulating data processing functions of the brain. ANNs have ahigh potential applicability in ecological systems due to their capacity to describehighly non-linear relationships among variables. In this sense, they represent promising computational tools to accurately model weed emergence and therefore, for prediction purposes to improve weed control. But they have not beenwidely used with this aim, and so far, they have been proven to be useful only for one weed species (Avena fatua). The objectives of the present work were to develop an ANN model for ripgut brome (Bromus diandrus) for emergence prediction and to compare their predictive capability against already developed non-linear regression (NLR) models. Thermal-time and hydro-time were used as independent input variablesfor developing bivariate models. The accumulated proportion of seedling emergence was the output variable. A total of 1610 input/output data pairs corresponding to three years of data collection in two different field trials wereused in this study. A total of 16 different scenarios or emergence data sets (differing in sowing dates and soil management) were modeledto compare the goodness of fit (RMSE) by the two approaches. The ANN developed had three layers: one input layer, one hidden layer with 2 neurons, and one output layer. Both procedures, ANN and NLR, were able to predict satisfactorily B. diandrus emergence patterns. However, the ANN improved the fitting accuracy in 11 of the 16 scenarios with RMSE estimates 46% lower compared to NLR models. These results confirm that ANNs are powerful tools for modeling weed emergence, thus they could help improve IWM decision support systems.Fil: Torra, Joel. Universidad de Lleida; EspañaFil: Royo Esnal, Aritz. Universidad de Lleida; EspañaFil: Chantre Balacca, Guillermo Ruben. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca. Centro de Recursos Naturales Renovables de la Zona Semiárida. Universidad Nacional del Sur. Centro de Recursos Naturales Renovables de la Zona Semiárida; Argentina. Universidad Nacional del Sur. Departamento de Agronomía; Argentina7th International Weed Science CongressPragaRepública ChecaInternational Weed Science SocietyCzech University of Life Sciences Prague2016info:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/conferenceObjectCongresoBookhttp://purl.org/coar/resource_type/c_5794info:ar-repo/semantics/documentoDeConferenciaapplication/pdfapplication/pdfapplication/pdfhttp://hdl.handle.net/11336/264213Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures; 7th International Weed Science Congress; Praga; República Checa; 2016; 27-27978-80-213-2648-4CONICET DigitalCONICETenginfo:eu-repo/semantics/altIdentifier/url/https://www.iwss.info/downloads/files/n6502b944c2d4a.pdfInternacionalinfo:eu-repo/semantics/openAccesshttps://creativecommons.org/licenses/by-nc-sa/2.5/ar/reponame:CONICET Digital (CONICET)instname:Consejo Nacional de Investigaciones Científicas y Técnicas2025-09-29T09:42:58Zoai:ri.conicet.gov.ar:11336/264213instacron:CONICETInstitucionalhttp://ri.conicet.gov.ar/Organismo científico-tecnológicoNo correspondehttp://ri.conicet.gov.ar/oai/requestdasensio@conicet.gov.ar; lcarlino@conicet.gov.arArgentinaNo correspondeNo correspondeNo correspondeopendoar:34982025-09-29 09:42:58.828CONICET Digital (CONICET) - Consejo Nacional de Investigaciones Científicas y Técnicasfalse
dc.title.none.fl_str_mv Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
title Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
spellingShingle Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
Torra, Joel
THERMAL TIME
HIDROTIME
GERMINATION
SIGMOIDAL REGRESSION
title_short Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
title_full Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
title_fullStr Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
title_full_unstemmed Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
title_sort Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures
dc.creator.none.fl_str_mv Torra, Joel
Royo Esnal, Aritz
Chantre Balacca, Guillermo Ruben
author Torra, Joel
author_facet Torra, Joel
Royo Esnal, Aritz
Chantre Balacca, Guillermo Ruben
author_role author
author2 Royo Esnal, Aritz
Chantre Balacca, Guillermo Ruben
author2_role author
author
dc.subject.none.fl_str_mv THERMAL TIME
HIDROTIME
GERMINATION
SIGMOIDAL REGRESSION
topic THERMAL TIME
HIDROTIME
GERMINATION
SIGMOIDAL REGRESSION
purl_subject.fl_str_mv https://purl.org/becyt/ford/4.1
https://purl.org/becyt/ford/4
https://purl.org/becyt/ford/1.2
https://purl.org/becyt/ford/1
dc.description.none.fl_txt_mv Artificial Neural Networks (ANNs) are machines with complexfunctional relations learnable with a limited amount of trainingdata emulating data processing functions of the brain. ANNs have ahigh potential applicability in ecological systems due to their capacity to describehighly non-linear relationships among variables. In this sense, they represent promising computational tools to accurately model weed emergence and therefore, for prediction purposes to improve weed control. But they have not beenwidely used with this aim, and so far, they have been proven to be useful only for one weed species (Avena fatua). The objectives of the present work were to develop an ANN model for ripgut brome (Bromus diandrus) for emergence prediction and to compare their predictive capability against already developed non-linear regression (NLR) models. Thermal-time and hydro-time were used as independent input variablesfor developing bivariate models. The accumulated proportion of seedling emergence was the output variable. A total of 1610 input/output data pairs corresponding to three years of data collection in two different field trials wereused in this study. A total of 16 different scenarios or emergence data sets (differing in sowing dates and soil management) were modeledto compare the goodness of fit (RMSE) by the two approaches. The ANN developed had three layers: one input layer, one hidden layer with 2 neurons, and one output layer. Both procedures, ANN and NLR, were able to predict satisfactorily B. diandrus emergence patterns. However, the ANN improved the fitting accuracy in 11 of the 16 scenarios with RMSE estimates 46% lower compared to NLR models. These results confirm that ANNs are powerful tools for modeling weed emergence, thus they could help improve IWM decision support systems.
Fil: Torra, Joel. Universidad de Lleida; España
Fil: Royo Esnal, Aritz. Universidad de Lleida; España
Fil: Chantre Balacca, Guillermo Ruben. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Científico Tecnológico Conicet - Bahía Blanca. Centro de Recursos Naturales Renovables de la Zona Semiárida. Universidad Nacional del Sur. Centro de Recursos Naturales Renovables de la Zona Semiárida; Argentina. Universidad Nacional del Sur. Departamento de Agronomía; Argentina
7th International Weed Science Congress
Praga
República Checa
International Weed Science Society
description Artificial Neural Networks (ANNs) are machines with complexfunctional relations learnable with a limited amount of trainingdata emulating data processing functions of the brain. ANNs have ahigh potential applicability in ecological systems due to their capacity to describehighly non-linear relationships among variables. In this sense, they represent promising computational tools to accurately model weed emergence and therefore, for prediction purposes to improve weed control. But they have not beenwidely used with this aim, and so far, they have been proven to be useful only for one weed species (Avena fatua). The objectives of the present work were to develop an ANN model for ripgut brome (Bromus diandrus) for emergence prediction and to compare their predictive capability against already developed non-linear regression (NLR) models. Thermal-time and hydro-time were used as independent input variablesfor developing bivariate models. The accumulated proportion of seedling emergence was the output variable. A total of 1610 input/output data pairs corresponding to three years of data collection in two different field trials wereused in this study. A total of 16 different scenarios or emergence data sets (differing in sowing dates and soil management) were modeledto compare the goodness of fit (RMSE) by the two approaches. The ANN developed had three layers: one input layer, one hidden layer with 2 neurons, and one output layer. Both procedures, ANN and NLR, were able to predict satisfactorily B. diandrus emergence patterns. However, the ANN improved the fitting accuracy in 11 of the 16 scenarios with RMSE estimates 46% lower compared to NLR models. These results confirm that ANNs are powerful tools for modeling weed emergence, thus they could help improve IWM decision support systems.
publishDate 2016
dc.date.none.fl_str_mv 2016
dc.type.none.fl_str_mv info:eu-repo/semantics/publishedVersion
info:eu-repo/semantics/conferenceObject
Congreso
Book
http://purl.org/coar/resource_type/c_5794
info:ar-repo/semantics/documentoDeConferencia
status_str publishedVersion
format conferenceObject
dc.identifier.none.fl_str_mv http://hdl.handle.net/11336/264213
Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures; 7th International Weed Science Congress; Praga; República Checa; 2016; 27-27
978-80-213-2648-4
CONICET Digital
CONICET
url http://hdl.handle.net/11336/264213
identifier_str_mv Modeling weed emergence: artifi cial neural networks versus non-linear regression procedures; 7th International Weed Science Congress; Praga; República Checa; 2016; 27-27
978-80-213-2648-4
CONICET Digital
CONICET
dc.language.none.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv info:eu-repo/semantics/altIdentifier/url/https://www.iwss.info/downloads/files/n6502b944c2d4a.pdf
dc.rights.none.fl_str_mv info:eu-repo/semantics/openAccess
https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
eu_rights_str_mv openAccess
rights_invalid_str_mv https://creativecommons.org/licenses/by-nc-sa/2.5/ar/
dc.format.none.fl_str_mv application/pdf
application/pdf
application/pdf
dc.coverage.none.fl_str_mv Internacional
dc.publisher.none.fl_str_mv Czech University of Life Sciences Prague
publisher.none.fl_str_mv Czech University of Life Sciences Prague
dc.source.none.fl_str_mv reponame:CONICET Digital (CONICET)
instname:Consejo Nacional de Investigaciones Científicas y Técnicas
reponame_str CONICET Digital (CONICET)
collection CONICET Digital (CONICET)
instname_str Consejo Nacional de Investigaciones Científicas y Técnicas
repository.name.fl_str_mv CONICET Digital (CONICET) - Consejo Nacional de Investigaciones Científicas y Técnicas
repository.mail.fl_str_mv dasensio@conicet.gov.ar; lcarlino@conicet.gov.ar
_version_ 1844613352093908992
score 13.070432