Predicting frost resistance of rubberized concrete for sustainable design: a comparative analysis of regression trees and stepwise regression models

Abstract

This study presents a machine learning-driven comparative analysis to predict frost resistance in rubberized concrete, focusing on the interpretability and performance of three models: Stepwise Linear Regression (SLR), Stepwise Polynomial Regression (SPR), and Classification and Regression Tree (CART). Leveraging historical experimental data, the methodology integrates rigorous preprocessing via Interquartile Range (IQR) outlier detection and tenfold cross-validation to ensure robustness. Key variables, including rubber content (0–20% fine aggregate replacement), water-cement ratio, and freeze–thaw cycles, were analyzed to balance durability and sustainability goals. The CART model offered interpretable decision rules (test R² = 97.01%) identifying critical thresholds like rubber content ≤ 15%, providing a clear guideline for maximizing durability. Comparatively, the study advances prior Artificial Neural Network (ANN)-based approaches by delivering mathematical transparency (SLR, SPR) and actionable insights (CART). These outputs directly guide mix design: the SPR equation enables precise prediction of frost resistance for specific combinations of mix parameters, while the CART rules establish safe application limits. This moves beyond prediction to offer practical tools for optimizing sustainable concrete mixes under frost exposure, prioritizing both accuracy and implementable insight.

Keywords

Rubberized concrete / Frost resistance / Stepwise regression / Sustainable design / Tenfold cross-validation

References

[1] Alam S, Gazder U. Classification and regression tree ( CART ) modelling for analysis of shear strength of FRP-RC members. Arab J Basic Appl Sci, 2021, 28(1): 397-405

[2] Bas YJ, Kakrasul JI, Ismail KS. Advanced predictive techniques for estimating compressive strength in recycled aggregate concrete: exploring interaction, quadratic models, ANN, and M5P across strength classes. Multiscale Multidiscip Model Exp Des, 2025, 8 122

[3] Breiman L (2017) Classification and regression trees. Routledge

[4] Bu C, et al.. Modification of rubberized concrete : a review. Buildings, 2022

[5] Cha GW, Hong WH, Lee KG (2015) A Study on Estimation of Demolition Waste Generation Index by using CART ( Classification and Regression Tree). Analysis Icsmme 90–93

[6] Chou J, Tsai C, Pham A, Lu Y. Machine learning in concrete strength simulations: multi-nation data analytics. Constr Build Mater, 2014, 73: 771-780

[7] Fu C, Li M. Frost resistance prediction for rubberized concrete based on artificial neural network. Discov Appl Sci, 2024

[8] Gao X, Yang J, Zhu H, Xu J. Estimation of rubberized concrete frost resistance using machine learning techniques. Constr Build Mater, 2023, 371 130778

[9] Hashem T, Kattoof I, Hassan N, Mohammed AY, Hussien ML (2024) Unlocking precision in hydraulic engineering : machine learning insights into labyrinth sluice gate discharge coef fi cients 26(11):2883–2901. https://doi.org/10.2166/hydro.2024.310

[10] Jiang L, Jing J, Zhang M, Yang S. Study on the effect of rubber content on the frost resistance of steel fiber reinforced rubber concrete. Sci Rep, 2024, 0123456789: 1-15

[11] Jie XU, Yao Z, Yang G, Han Q. Research on crumb rubber concrete: from a multi-scale review. Constr Build Mater, 2020, 232 117282

[12] Kattoof I, Ahmed H, Abdulhadi M, Hussien ML (2024) Prediction of concrete strength by hierarchical stepwise regression using ultrasonic pulse velocity and Schmidt rebound hammer. Innov Infrastruct Solut 1–11. https://doi.org/10.1007/s41062-024-01766-1

[13] Kattoof I, Ehsan H, Salman E, Hussien ML, Mohammed AY, Nadir W. Optimizing compressive strength of foamed concrete using stepwise regression. Discov Appl Sci, 2025

[14] Kattoof I, Mousa A, Hussien ML. Harnessing machine learning for accurate estimation of compressive strength of high-performance self-compacting concrete from non-destructive tests : A comparative study. Constr. Build. Mater., 2024, 451: 138779

[15] Kattoof I, Muna H, Mousa MAA, Mohammed A (2025) Harnessing machine learning for accurate estimation of concrete strength using non-destructive tests : a comparative study. Multiscale Multidiscip Model Exp Des 123. https://doi.org/10.1007/s41939-024-00605-1

[16] Kattoof I, Wissam H, Mustafa N, Ahmed SS, Mohammed YM. Machine learning - based prediction of shear strength in interior beam - column joints. Discov Appl Sci, 2025

[17] Kattoof I, Wissam H, Mustafa N, Mohammed SS (2024) Prediction of high-performance concrete strength using machine learning with hierarchical regression. Multiscale Multidiscip Model Exp. Des 2009. https://doi.org/10.1007/s41939-024-00467-7

[18] Khan MA et al (2021) Geopolymer Concrete Compressive Strength via Arti fi cial Neural Network , Adaptive Neuro Fuzzy Interface System , and Gene Expression Programming With K-Fold Cross Validation. 8:1-19. https://doi.org/10.3389/fmats.2021.621163

[19] Kohavi R (1995) A Study of Cross-Validation and Bootstrap for Accuracy Estimation and Model Selection

[20] Li G, Garrick G, Eggers J, Abadie C, Stubblefield MA, Pang S. Waste tire fiber modified concrete. Compos Part B Eng, 2004, 35: 305-312

[21] Loh W (n.d.) Classification and Regression Trees 1–17

[22] Oikonomou N, Mavridou S (2009) Cement & Concrete Composites Improvement of chloride ion penetration resistance in cement mortars modified with rubber from worn automobile tires. 31:403–407. https://doi.org/10.1016/j.cemconcomp.2009.04.004

[23] Paine KA, Dhir RK, Moroney R, Kopasakis K (2002) Use of crumb rubber to achieve freeze thaw resisting concrete. In Proceedings of the international conference on concrete for extreme conditions, University of Dundee, Scotland, UK 9(11)

[24] Pongsopha P, Sukontasukkul P, Zhang H, Limkatanyu S (2022) Results in Engineering Thermal and acoustic properties of sustainable structural lightweight aggregate rubberized concrete. Results Eng 13:100333. https://doi.org/10.1016/j.rineng.2022.100333

[25] Radiation S, Duration S, Models E (2020) “Sustainable Energy Technologies and Assessments, 40:100768 https://doi.org/10.1016/j.seta.2020.100768

[26] Sanjeev C, Dash K, Kumar A, Dehuri S, Ghosh A (2023) An outliers detection and elimination framework in classification task of data mining. Decis Anal J 6:100164. https://doi.org/10.1016/j.dajour.2023.100164

[27] Savas BZ, Ahmad S, Fedroff D (1997) Freeze-Thaw Durability of Concrete with Ground Waste Tire Rubber. Transp Res Rec 1574(1):89–88. https://doi.org/10.3141/1574-11

[28] Segre N, Joekes I. Use of tire rubber particles as addition to cement paste. Cem Concr Res, 2000, 30: 1421-1425

[29] Statistics C et al (2017) A simple more general boxplot method for identifying outliers A simple more general boxplot method for identifying outliers. https://doi.org/10.1016/j.csda.2003.10.012

[30] Struct JC, et al.. Experimental Analysis and Evaluation of the Compressive Strength of Rubberized Concrete During Freeze – Thaw Cycles. Int J Concr Struct Mater, 2023

[31] Taylor KE (2001) In a single diagram 106:7183–7192

[32] Tohidi S, Sharifi Y. Empirical modeling of distortional buckling strength of half-through bridge girders via stepwise regression method. Adv Struct Eng, 2015, 18(9): 1383-1398

[33] Turatsinze A, Bonnet S. Potential of rubber aggregates to modify properties of cement based-mortars : improvement in cracking shrinkage resistance. Constr Build Mater, 2007, 21(1): 176-181

[34] Zhu X, Miao C, Liu J, Hong J. Influence of crumb rubber on frost resistance of concrete and effect mechanism. Procedia Eng, 2012

[35] Zhang B, Poon CS (2017) SC. J Clean Prod. https://doi.org/10.1016/j.jclepro.2017.11.044

RIGHTS & PERMISSIONS

The Author(s)