Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification

Detalhes bibliográficos
Autor(a) principal: Buarque, Thiago Matheus Torres
Data de Publicação: 2022
Outros Autores: Marinho, Matheus Barreto Lins, Bernardino Junior, Francisco Madeiro
Tipo de documento: Artigo
Idioma: por
Título da fonte: Research, Society and Development
Texto Completo: https://rsdjournal.org/index.php/rsd/article/view/26984
Resumo: Non-functional requirements (NFRs) play an important role in the Software Engineering (SE) area, being associated with the construction, operation, and maintenance of a quality application. The success of the RNF classification manual task depends on the knowledge and experience of the requirements engineer and is time-consuming. Works have been developed aiming at the application of machine learning algorithms to automatically classify RNFs, a scenario in which the hyperparameters of the classifier model must be chosen. In this work, the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) algorithm are used to find hyperparameters of the Multilayer Neural Perceptron Network (MLP), with the objective of classifying NFRs present in the PROMISE_exp dataset. The GA found a combination of hyperparameters that gave an F1 of 0.6349, while the PSO found a combination that got 0.6426 of F1.
id UNIFEI_5fb552774cc43feeb087bc7d485966a0
oai_identifier_str oai:ojs.pkp.sfu.ca:article/26984
network_acronym_str UNIFEI
network_name_str Research, Society and Development
repository_id_str
spelling Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements ClassificationAlgoritmo Genético y PSO Aplicados a la Elección de los Hiperparámetros de una Red Neuronal MLP para la Clasificación de Requisitos no FuncionalesAlgoritmo Genético e PSO Aplicados à Escolha dos Hiperparâmetros de uma Rede Neural MLP para Classificação de Requisitos Não-FuncionaisClasificaciónRequisitos de softwareAlgoritmo genéticoPSOPerceptrón multicapa.ClassificaçãoRequisitos de softwareAlgoritmo genéticoPSOPerceptron multicamadas.ClassificationSoftware requirementsGenetic AlgorithmPSOMultilayer Perceptron.Non-functional requirements (NFRs) play an important role in the Software Engineering (SE) area, being associated with the construction, operation, and maintenance of a quality application. The success of the RNF classification manual task depends on the knowledge and experience of the requirements engineer and is time-consuming. Works have been developed aiming at the application of machine learning algorithms to automatically classify RNFs, a scenario in which the hyperparameters of the classifier model must be chosen. In this work, the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) algorithm are used to find hyperparameters of the Multilayer Neural Perceptron Network (MLP), with the objective of classifying NFRs present in the PROMISE_exp dataset. The GA found a combination of hyperparameters that gave an F1 of 0.6349, while the PSO found a combination that got 0.6426 of F1.Los requisitos no funcionales (RNFs) juegan un papel importante en el área de Ingeniería de Software (IS), estando asociados con la construcción, operación y mantenimiento de una aplicación de calidad. El éxito de la tarea del manual de clasificación RNF depende del conocimiento y la experiencia del ingeniero de requisitos y requiere mucho tiempo. Se han desarrollado trabajos encaminados a la aplicación de algoritmos de aprendizaje automático para clasificar automáticamente los RNF, escenario en el que se deben elegir los hiperparámetros del modelo clasificador. En este trabajo, el Algoritmo Genético (GA, Genetic Algorithm) y el algoritmo de Optimización de Enjambre de Partículas (PSO, Particle Swarm Optimization) se utilizan para encontrar hiperparámetros de la Red de Perceptrón Neural Multicapa (MLP, Multilayer Perceptron), con el objetivo de clasificar los RNF presentes en el PROMISE_exp conjunto de datos. La GA encontró una combinación de hiperparámetros que dio un F1 de 0.6349, mientras que el PSO encontró una combinación que obtuvo 0.6426 de F1.Os requisitos não-funcionais (RNFs) desempenham um papel importante na área de Engenharia de Software (ES), estando associados à construção, ao funcionamento e à manutenção de uma aplicação de qualidade. O sucesso da tarefa manual de classificação de RNFs depende do conhecimento e da experiência do engenheiro de requisitos, além de demandar tempo. Trabalhos têm sido desenvolvidos visando a aplicação de algoritmos de aprendizagem de máquina para classificar automaticamente RNFs, cenário em que devem ser escolhidos os hiperparâmetros do modelo classificador. Neste trabalho, o Algoritmo Genético (GA, Genetic Algorithm) e o algoritmo de Otimização por Enxame de Partículas (PSO, Particle Swarm Optimization) são utilizados para encontrar hiperparâmetros da rede Neural Perceptron Multicamada (MLP, Multilayer Perceptron), com o objetivo de classificar RNFs presentes no conjunto de dados PROMISE_exp. O GA encontrou uma combinação de hiperparâmetros que resultou em F1 de 0,6349, enquanto o PSO encontrou uma combinação que obteve 0,6426 de F1.Research, Society and Development2022-03-07info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersionapplication/pdfhttps://rsdjournal.org/index.php/rsd/article/view/2698410.33448/rsd-v11i3.26984Research, Society and Development; Vol. 11 No. 3; e55411326984Research, Society and Development; Vol. 11 Núm. 3; e55411326984Research, Society and Development; v. 11 n. 3; e554113269842525-3409reponame:Research, Society and Developmentinstname:Universidade Federal de Itajubá (UNIFEI)instacron:UNIFEIporhttps://rsdjournal.org/index.php/rsd/article/view/26984/23539Copyright (c) 2022 Thiago Matheus Torres Buarque; Matheus Barreto Lins Marinho; Francisco Madeiro Bernardino Juniorhttps://creativecommons.org/licenses/by/4.0info:eu-repo/semantics/openAccessBuarque, Thiago Matheus TorresMarinho, Matheus Barreto LinsBernardino Junior, Francisco Madeiro2022-03-09T13:44:38Zoai:ojs.pkp.sfu.ca:article/26984Revistahttps://rsdjournal.org/index.php/rsd/indexPUBhttps://rsdjournal.org/index.php/rsd/oairsd.articles@gmail.com2525-34092525-3409opendoar:2024-01-17T09:44:51.833187Research, Society and Development - Universidade Federal de Itajubá (UNIFEI)false
dc.title.none.fl_str_mv Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
Algoritmo Genético y PSO Aplicados a la Elección de los Hiperparámetros de una Red Neuronal MLP para la Clasificación de Requisitos no Funcionales
Algoritmo Genético e PSO Aplicados à Escolha dos Hiperparâmetros de uma Rede Neural MLP para Classificação de Requisitos Não-Funcionais
title Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
spellingShingle Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
Buarque, Thiago Matheus Torres
Clasificación
Requisitos de software
Algoritmo genético
PSO
Perceptrón multicapa.
Classificação
Requisitos de software
Algoritmo genético
PSO
Perceptron multicamadas.
Classification
Software requirements
Genetic Algorithm
PSO
Multilayer Perceptron.
title_short Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
title_full Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
title_fullStr Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
title_full_unstemmed Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
title_sort Genetic Algorithm and PSO Applied to the Choice of Hyperparameters of an MLP Neural Network for Non-Functional Requirements Classification
author Buarque, Thiago Matheus Torres
author_facet Buarque, Thiago Matheus Torres
Marinho, Matheus Barreto Lins
Bernardino Junior, Francisco Madeiro
author_role author
author2 Marinho, Matheus Barreto Lins
Bernardino Junior, Francisco Madeiro
author2_role author
author
dc.contributor.author.fl_str_mv Buarque, Thiago Matheus Torres
Marinho, Matheus Barreto Lins
Bernardino Junior, Francisco Madeiro
dc.subject.por.fl_str_mv Clasificación
Requisitos de software
Algoritmo genético
PSO
Perceptrón multicapa.
Classificação
Requisitos de software
Algoritmo genético
PSO
Perceptron multicamadas.
Classification
Software requirements
Genetic Algorithm
PSO
Multilayer Perceptron.
topic Clasificación
Requisitos de software
Algoritmo genético
PSO
Perceptrón multicapa.
Classificação
Requisitos de software
Algoritmo genético
PSO
Perceptron multicamadas.
Classification
Software requirements
Genetic Algorithm
PSO
Multilayer Perceptron.
description Non-functional requirements (NFRs) play an important role in the Software Engineering (SE) area, being associated with the construction, operation, and maintenance of a quality application. The success of the RNF classification manual task depends on the knowledge and experience of the requirements engineer and is time-consuming. Works have been developed aiming at the application of machine learning algorithms to automatically classify RNFs, a scenario in which the hyperparameters of the classifier model must be chosen. In this work, the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO) algorithm are used to find hyperparameters of the Multilayer Neural Perceptron Network (MLP), with the objective of classifying NFRs present in the PROMISE_exp dataset. The GA found a combination of hyperparameters that gave an F1 of 0.6349, while the PSO found a combination that got 0.6426 of F1.
publishDate 2022
dc.date.none.fl_str_mv 2022-03-07
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://rsdjournal.org/index.php/rsd/article/view/26984
10.33448/rsd-v11i3.26984
url https://rsdjournal.org/index.php/rsd/article/view/26984
identifier_str_mv 10.33448/rsd-v11i3.26984
dc.language.iso.fl_str_mv por
language por
dc.relation.none.fl_str_mv https://rsdjournal.org/index.php/rsd/article/view/26984/23539
dc.rights.driver.fl_str_mv https://creativecommons.org/licenses/by/4.0
info:eu-repo/semantics/openAccess
rights_invalid_str_mv https://creativecommons.org/licenses/by/4.0
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Research, Society and Development
publisher.none.fl_str_mv Research, Society and Development
dc.source.none.fl_str_mv Research, Society and Development; Vol. 11 No. 3; e55411326984
Research, Society and Development; Vol. 11 Núm. 3; e55411326984
Research, Society and Development; v. 11 n. 3; e55411326984
2525-3409
reponame:Research, Society and Development
instname:Universidade Federal de Itajubá (UNIFEI)
instacron:UNIFEI
instname_str Universidade Federal de Itajubá (UNIFEI)
instacron_str UNIFEI
institution UNIFEI
reponame_str Research, Society and Development
collection Research, Society and Development
repository.name.fl_str_mv Research, Society and Development - Universidade Federal de Itajubá (UNIFEI)
repository.mail.fl_str_mv rsd.articles@gmail.com
_version_ 1797052793367822336

Registros relacionados