Non-linear and mixed regression models in predicting sustainable concrete strength

Ruoyu Jin, Qian Chen, Alfred Soboyejo

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Most previous research adopting the regression analysis to capture the relationship between concrete properties and mixture-design-related variables was based on the linear approach with limited accuracy. This study applies non-linear and mixed regression analysis to model properties of environmentally friendly concrete based on a comprehensive set of variables containing alternative or waste materials. It was found that best-fit non-linear and mixed models achieved similar accuracies and superior R2 values compared to the linear approach when using both the numerical and relative input methods. Individual materials’ effects on concrete strength were statistically quantified at different curing ages using the best-fit models.
Original languageEnglish
Pages (from-to)142-152
Number of pages11
JournalConstruction and Building Materials
Volume170
DOIs
Publication statusPublished - 23 Mar 2018

Fingerprint

Concretes
Regression analysis
Curing

Bibliographical note

© 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/

Cite this

@article{ff4336f2d8e04b3b91ef5a6f54c77f4d,
title = "Non-linear and mixed regression models in predicting sustainable concrete strength",
abstract = "Most previous research adopting the regression analysis to capture the relationship between concrete properties and mixture-design-related variables was based on the linear approach with limited accuracy. This study applies non-linear and mixed regression analysis to model properties of environmentally friendly concrete based on a comprehensive set of variables containing alternative or waste materials. It was found that best-fit non-linear and mixed models achieved similar accuracies and superior R2 values compared to the linear approach when using both the numerical and relative input methods. Individual materials’ effects on concrete strength were statistically quantified at different curing ages using the best-fit models.",
author = "Ruoyu Jin and Qian Chen and Alfred Soboyejo",
note = "{\circledC} 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/",
year = "2018",
month = "3",
day = "23",
doi = "10.1016/j.conbuildmat.2018.03.063",
language = "English",
volume = "170",
pages = "142--152",
journal = "Construction and Building Materials",
issn = "0950-0618",

}

Non-linear and mixed regression models in predicting sustainable concrete strength. / Jin, Ruoyu; Chen, Qian; Soboyejo, Alfred.

In: Construction and Building Materials, Vol. 170, 23.03.2018, p. 142-152.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Non-linear and mixed regression models in predicting sustainable concrete strength

AU - Jin, Ruoyu

AU - Chen, Qian

AU - Soboyejo, Alfred

N1 - © 2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/

PY - 2018/3/23

Y1 - 2018/3/23

N2 - Most previous research adopting the regression analysis to capture the relationship between concrete properties and mixture-design-related variables was based on the linear approach with limited accuracy. This study applies non-linear and mixed regression analysis to model properties of environmentally friendly concrete based on a comprehensive set of variables containing alternative or waste materials. It was found that best-fit non-linear and mixed models achieved similar accuracies and superior R2 values compared to the linear approach when using both the numerical and relative input methods. Individual materials’ effects on concrete strength were statistically quantified at different curing ages using the best-fit models.

AB - Most previous research adopting the regression analysis to capture the relationship between concrete properties and mixture-design-related variables was based on the linear approach with limited accuracy. This study applies non-linear and mixed regression analysis to model properties of environmentally friendly concrete based on a comprehensive set of variables containing alternative or waste materials. It was found that best-fit non-linear and mixed models achieved similar accuracies and superior R2 values compared to the linear approach when using both the numerical and relative input methods. Individual materials’ effects on concrete strength were statistically quantified at different curing ages using the best-fit models.

U2 - 10.1016/j.conbuildmat.2018.03.063

DO - 10.1016/j.conbuildmat.2018.03.063

M3 - Article

VL - 170

SP - 142

EP - 152

JO - Construction and Building Materials

JF - Construction and Building Materials

SN - 0950-0618

ER -