Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model

Concrete carbonation is a critical factor influencing the durability and longevity of concrete structures, particularly in urban environments. Traditional methods for predicting the anti-carbonation performance (ACP) of concrete often lack precision and fail to account for complex interactions betw...

Full description

Saved in:
Bibliographic Details
Main Authors: Sivaraja M., Swaminathen A. N., Kuttimarks M. S., Rajprasad J., Sakthivel M., Rex J.
Format: Article
Language:English
Published: Vilnius Gediminas Technical University 2025-05-01
Series:Journal of Environmental Engineering and Landscape Management
Subjects:
Online Access:https://journals.vilniustech.lt/index.php/JEELM/article/view/23568
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849730475157880832
author Sivaraja M.
Swaminathen A. N.
Kuttimarks M. S.
Rajprasad J.
Sakthivel M.
Rex J.
author_facet Sivaraja M.
Swaminathen A. N.
Kuttimarks M. S.
Rajprasad J.
Sakthivel M.
Rex J.
author_sort Sivaraja M.
collection DOAJ
description Concrete carbonation is a critical factor influencing the durability and longevity of concrete structures, particularly in urban environments. Traditional methods for predicting the anti-carbonation performance (ACP) of concrete often lack precision and fail to account for complex interactions between influencing factors. In this study, a novel hybrid model combining random forest (RF) regression with a least squares support vector machine (LSSVM) is proposed to enhance the accuracy of ACP predictions. The RF regression is utilized for feature selection, identifying the most significant factors affecting ACP and optimizing the input features for the LSSVM model. Our hybrid model is validated against a comprehensive dataset, demonstrating superior performance in predicting concrete carbonation resistance compared to conventional methods. Quantitative results show that the RF-LSSVM model achieves a root mean square error (RMSE) of 5e–5 and a high coefficient of determination (R-squared) of 0.999, indicating robust predictive capability and accuracy. The main novelty of this work lies in the integration of RF and LSSVM to create a robust, accurate, and practical tool for assessing the durability of concrete structures.
format Article
id doaj-art-e47a784fbbe94cc497e21569c98178b6
institution DOAJ
issn 1648-6897
1822-4199
language English
publishDate 2025-05-01
publisher Vilnius Gediminas Technical University
record_format Article
series Journal of Environmental Engineering and Landscape Management
spelling doaj-art-e47a784fbbe94cc497e21569c98178b62025-08-20T03:08:51ZengVilnius Gediminas Technical UniversityJournal of Environmental Engineering and Landscape Management1648-68971822-41992025-05-0133210.3846/jeelm.2025.23568Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine modelSivaraja M.0Swaminathen A. N.1Kuttimarks M. S.2Rajprasad J.3Sakthivel M.4Rex J.5Department of Civil Engineering, Nehru Institute of Technology, Coimbatore, 641105 Tamilnadu, IndiaDepartment of Civil Engineering, Adi Shankara Institute of Engineering and Technology, Kalady, KeralaDepartment of Civil Engineering, Shivajirao S. Jondhle College of Engineering & Technology, Thane, MaharastraDepartment of Civil Engineering, College of Engineering and Technology, SRM Institute of Science and Technology, 603203 Kattakulathur, Tamilnadu, IndiaDepartment of Civil Engineering, Kongunadu College of Engineering and Technology, 621215 Trichy, IndiaDepartment of Civil Engineering, CK College of Engineering & Technology, 607003 Cuddalore, Tamilnadu, India Concrete carbonation is a critical factor influencing the durability and longevity of concrete structures, particularly in urban environments. Traditional methods for predicting the anti-carbonation performance (ACP) of concrete often lack precision and fail to account for complex interactions between influencing factors. In this study, a novel hybrid model combining random forest (RF) regression with a least squares support vector machine (LSSVM) is proposed to enhance the accuracy of ACP predictions. The RF regression is utilized for feature selection, identifying the most significant factors affecting ACP and optimizing the input features for the LSSVM model. Our hybrid model is validated against a comprehensive dataset, demonstrating superior performance in predicting concrete carbonation resistance compared to conventional methods. Quantitative results show that the RF-LSSVM model achieves a root mean square error (RMSE) of 5e–5 and a high coefficient of determination (R-squared) of 0.999, indicating robust predictive capability and accuracy. The main novelty of this work lies in the integration of RF and LSSVM to create a robust, accurate, and practical tool for assessing the durability of concrete structures. https://journals.vilniustech.lt/index.php/JEELM/article/view/23568carbonation resistanceleast squares support vector machine (LSSVM) modelsafety assessmentrandom forest (RF)concrete
spellingShingle Sivaraja M.
Swaminathen A. N.
Kuttimarks M. S.
Rajprasad J.
Sakthivel M.
Rex J.
Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
Journal of Environmental Engineering and Landscape Management
carbonation resistance
least squares support vector machine (LSSVM) model
safety assessment
random forest (RF)
concrete
title Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
title_full Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
title_fullStr Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
title_full_unstemmed Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
title_short Prediction of the anti-carbonation performance of concrete based on random forest – least squares support vector machine model
title_sort prediction of the anti carbonation performance of concrete based on random forest least squares support vector machine model
topic carbonation resistance
least squares support vector machine (LSSVM) model
safety assessment
random forest (RF)
concrete
url https://journals.vilniustech.lt/index.php/JEELM/article/view/23568
work_keys_str_mv AT sivarajam predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel
AT swaminathenan predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel
AT kuttimarksms predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel
AT rajprasadj predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel
AT sakthivelm predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel
AT rexj predictionoftheanticarbonationperformanceofconcretebasedonrandomforestleastsquaressupportvectormachinemodel