Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques

Detalhes bibliográficos
Autor(a) principal: Mateus, Miguel Santos Reis
Data de Publicação: 2023
Tipo de documento: Dissertação
Idioma: eng
Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
Texto Completo: http://hdl.handle.net/10362/148541
Resumo: Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
id RCAP_05c56e7604b2637cf6b57dc8edb1137e
oai_identifier_str oai:run.unl.pt:10362/148541
network_acronym_str RCAP
network_name_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository_id_str 7160
spelling Prediction of high-performance concrete compressive strength through a comparison of machine learning techniquesMachine LearningHigh-performance concreteCompressive strengthGradient BoostModelingDomínio/Área Científica::Ciências Naturais::Ciências da Computação e da InformaçãoDissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data ScienceHigh-performance concrete (HPC) is a highly complex composite material whose characteristics are extremely difficult to model. One of those characteristics is the concrete compressive strength, a nonlinear function of the same ingredients that compose HPC: cement, fly ash, blast furnace slag, water, superplasticizer, age, and coarse and fine aggregates. Research has shown time and time again that concrete strength is not determined just by the water-to-cement ratio, which was for years the go to metric. In addition, traditional methods that attempt to model HPC, such as regression analysis, do not provide sufficient prediction power due to nonlinear proprieties of the mixture. Therefore, this study attempts to optimize the prediction and modeling of the compressive strength of HPC by analyzing seven different machine learning (ML) algorithms: three regularization algorithms (Lasso, Ridge and Elastic Net), three ensemble algorithms (Random Forest, Gradient Boost and AdaBoost), and Artificial Neural Networks. All techniques were built and tested with a dataset composed of data from 17 different concrete strength test laboratories, under the same experimental conditions, which enabled a fair comparison amongst them and between different previous studies in the field. Feature importance analysis and outlier analysis were also performed, and all models were subject to a Wilcoxon Signed-Ranks Test to ensure statistically significant results. The final results show that the more complex ML algorithms provided greater accuracy than the regularization techniques, with Gradient Boost being the superior model amongst them, providing more accurate predictions than the sate-of-the-art. Better results were achieved using all variables and without removing outlier observations.Vanneschi, LeonardoRUNMateus, Miguel Santos Reis2023-02-02T14:25:10Z2023-01-232023-01-23T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10362/148541TID:203211715enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-03-11T05:30:10Zoai:run.unl.pt:10362/148541Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T03:53:24.535581Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse
dc.title.none.fl_str_mv Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
title Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
spellingShingle Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
Mateus, Miguel Santos Reis
Machine Learning
High-performance concrete
Compressive strength
Gradient Boost
Modeling
Domínio/Área Científica::Ciências Naturais::Ciências da Computação e da Informação
title_short Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
title_full Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
title_fullStr Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
title_full_unstemmed Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
title_sort Prediction of high-performance concrete compressive strength through a comparison of machine learning techniques
author Mateus, Miguel Santos Reis
author_facet Mateus, Miguel Santos Reis
author_role author
dc.contributor.none.fl_str_mv Vanneschi, Leonardo
RUN
dc.contributor.author.fl_str_mv Mateus, Miguel Santos Reis
dc.subject.por.fl_str_mv Machine Learning
High-performance concrete
Compressive strength
Gradient Boost
Modeling
Domínio/Área Científica::Ciências Naturais::Ciências da Computação e da Informação
topic Machine Learning
High-performance concrete
Compressive strength
Gradient Boost
Modeling
Domínio/Área Científica::Ciências Naturais::Ciências da Computação e da Informação
description Dissertation presented as the partial requirement for obtaining a Master's degree in Data Science and Advanced Analytics, specialization in Data Science
publishDate 2023
dc.date.none.fl_str_mv 2023-02-02T14:25:10Z
2023-01-23
2023-01-23T00:00:00Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://hdl.handle.net/10362/148541
TID:203211715
url http://hdl.handle.net/10362/148541
identifier_str_mv TID:203211715
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.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron:RCAAP
instname_str Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron_str RCAAP
institution RCAAP
reponame_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
collection Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository.name.fl_str_mv Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
repository.mail.fl_str_mv
_version_ 1799138124569247744

Registros relacionados