ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE

Detalhes bibliográficos
Autor(a) principal: Clemente,Marco A. J.
Data de Publicação: 2022
Outros Autores: Silva,Heloisa H. P., Campos,Júlia W., Mantovani,Ana C. G., Borsato,Dionisio
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Química Nova (Online)
Texto Completo: http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-40422022000800935
Resumo: Biodiesel is capable of replacing diesel because it has similar physicochemical properties, but this biofuel is susceptible to oxidation, which makes the application of antioxidant substances necessary. For this study, alcoholic extracts of senna leaves, hibiscus flowers, and blackberry were used. Biodiesel samples were submitted to physicochemical analysis to evaluate interference in the volume of these alcoholic extracts with antioxidant properties. The data obtained were processed using the neural network of the multilayer perceptron type (MLP). For the network’s training, 200 epochs were used. The samples were randomly divided into three groups, with 70% used for training, 15% for testing, and 15% for validation. The type of extract was considered as a categorical variable, the extract volume as a target variable, and the other ones as input variables. Among the 200 networks trained, with 5 to 20 hidden layers, the 5 with the best performance were highlighted. The Tukey test applied to the means showed no significant difference at the 5% level, between the value of the added volume and the means value predicted by the networks. The sensitive analysis showed that the most important input variable for the construction of the model was the type of extract.
id SBQ-3_8ab73a43126e08fd092c550c9753daa9
oai_identifier_str oai:scielo:S0100-40422022000800935
network_acronym_str SBQ-3
network_name_str Química Nova (Online)
repository_id_str
spelling ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPEmathematical modelingsenna leaveshibiscus flowersblackberry fruits.Biodiesel is capable of replacing diesel because it has similar physicochemical properties, but this biofuel is susceptible to oxidation, which makes the application of antioxidant substances necessary. For this study, alcoholic extracts of senna leaves, hibiscus flowers, and blackberry were used. Biodiesel samples were submitted to physicochemical analysis to evaluate interference in the volume of these alcoholic extracts with antioxidant properties. The data obtained were processed using the neural network of the multilayer perceptron type (MLP). For the network’s training, 200 epochs were used. The samples were randomly divided into three groups, with 70% used for training, 15% for testing, and 15% for validation. The type of extract was considered as a categorical variable, the extract volume as a target variable, and the other ones as input variables. Among the 200 networks trained, with 5 to 20 hidden layers, the 5 with the best performance were highlighted. The Tukey test applied to the means showed no significant difference at the 5% level, between the value of the added volume and the means value predicted by the networks. The sensitive analysis showed that the most important input variable for the construction of the model was the type of extract.Sociedade Brasileira de Química2022-08-01info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersiontext/htmlhttp://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-40422022000800935Química Nova v.45 n.8 2022reponame:Química Nova (Online)instname:Sociedade Brasileira de Química (SBQ)instacron:SBQ10.21577/0100-4042.20170902info:eu-repo/semantics/openAccessClemente,Marco A. J.Silva,Heloisa H. P.Campos,Júlia W.Mantovani,Ana C. G.Borsato,Dionisioeng2022-10-24T00:00:00Zoai:scielo:S0100-40422022000800935Revistahttps://www.scielo.br/j/qn/ONGhttps://old.scielo.br/oai/scielo-oai.phpquimicanova@sbq.org.br1678-70640100-4042opendoar:2022-10-24T00:00Química Nova (Online) - Sociedade Brasileira de Química (SBQ)false
dc.title.none.fl_str_mv ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
title ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
spellingShingle ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
Clemente,Marco A. J.
mathematical modeling
senna leaves
hibiscus flowers
blackberry fruits.
title_short ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
title_full ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
title_fullStr ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
title_full_unstemmed ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
title_sort ADDITION OF NATURAL EXTRACTS WITH ANTIOXIDANT PROPERTIES IN BIODIESEL: ANALYSIS BY NEURAL NETWORKS OF THE MULTILAYER PERCEPTRON TYPE
author Clemente,Marco A. J.
author_facet Clemente,Marco A. J.
Silva,Heloisa H. P.
Campos,Júlia W.
Mantovani,Ana C. G.
Borsato,Dionisio
author_role author
author2 Silva,Heloisa H. P.
Campos,Júlia W.
Mantovani,Ana C. G.
Borsato,Dionisio
author2_role author
author
author
author
dc.contributor.author.fl_str_mv Clemente,Marco A. J.
Silva,Heloisa H. P.
Campos,Júlia W.
Mantovani,Ana C. G.
Borsato,Dionisio
dc.subject.por.fl_str_mv mathematical modeling
senna leaves
hibiscus flowers
blackberry fruits.
topic mathematical modeling
senna leaves
hibiscus flowers
blackberry fruits.
description Biodiesel is capable of replacing diesel because it has similar physicochemical properties, but this biofuel is susceptible to oxidation, which makes the application of antioxidant substances necessary. For this study, alcoholic extracts of senna leaves, hibiscus flowers, and blackberry were used. Biodiesel samples were submitted to physicochemical analysis to evaluate interference in the volume of these alcoholic extracts with antioxidant properties. The data obtained were processed using the neural network of the multilayer perceptron type (MLP). For the network’s training, 200 epochs were used. The samples were randomly divided into three groups, with 70% used for training, 15% for testing, and 15% for validation. The type of extract was considered as a categorical variable, the extract volume as a target variable, and the other ones as input variables. Among the 200 networks trained, with 5 to 20 hidden layers, the 5 with the best performance were highlighted. The Tukey test applied to the means showed no significant difference at the 5% level, between the value of the added volume and the means value predicted by the networks. The sensitive analysis showed that the most important input variable for the construction of the model was the type of extract.
publishDate 2022
dc.date.none.fl_str_mv 2022-08-01
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-40422022000800935
url http://old.scielo.br/scielo.php?script=sci_arttext&pid=S0100-40422022000800935
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv 10.21577/0100-4042.20170902
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv text/html
dc.publisher.none.fl_str_mv Sociedade Brasileira de Química
publisher.none.fl_str_mv Sociedade Brasileira de Química
dc.source.none.fl_str_mv Química Nova v.45 n.8 2022
reponame:Química Nova (Online)
instname:Sociedade Brasileira de Química (SBQ)
instacron:SBQ
instname_str Sociedade Brasileira de Química (SBQ)
instacron_str SBQ
institution SBQ
reponame_str Química Nova (Online)
collection Química Nova (Online)
repository.name.fl_str_mv Química Nova (Online) - Sociedade Brasileira de Química (SBQ)
repository.mail.fl_str_mv quimicanova@sbq.org.br
_version_ 1750318121940418560