Jul 2026· International Conference on Machine Learning and Embedded Systems· Vol 14295, pp. 142950L - 142950L-6· 0 citations· 7 references
Engineering
TL;DR
Results suggest that, within the present five-fold cross-validation setting and limited-sample dataset, RBF kernel ridge regression captures the nonlinear relationships more effectively than conventional linear models; however, broader generalization should be verified using larger datasets and additional validation.
Abstract
Accurate evaluation of the mechanical performance of modified concrete is important for mixture design and engineering application. This study developed machine-learning models to predict the compressive strength (CS) and flexural strength (FS) of rubber-polymer modified concrete using 50 experimental samples. Nine variables-cement, water, fine aggregate, coarse aggregate, superplasticizer, rubber content, polymer content, polymer solid content, and rubber particle size-were used as input features. Four regression models, namely linear regression, ridge regression, RBF kernel ridge regression, and bagged linear regression, were trained for two independent prediction tasks. Model performance was assessed by fivefold cross-validation using RMSE, MAE, and R2. Within a unified framework integrating data characterization, model comparison, and post-hoc interpretation, RBF kernel ridge regression achieved the best overall performance for both targets. For CS prediction, the RMSE, MAE, and R2 were 6.656, 4.921, and 0.824, respectively; for FS prediction, the corresponding values were 1.063, 0.644, and 0.802. Correlation analysis and ridge-coefficient interpretation showed that CS was more sensitive to coarse aggregate, water, superplasticizer dosage, and polymer solid content, whereas FS was more strongly influenced by fine aggregate, water, polymer solid content, and cement. These results suggest that, within the present five-fold cross-validation setting and limited-sample dataset, RBF kernel ridge regression captures the nonlinear relationships more effectively than conventional linear models; however, broader generalization should be verified using larger datasets and additional validation.
The use of supplementary cementitious materials such as fly ash can reduce environmental impacts and improve the sustainability of concrete construction. However, the nonlinear interactions among mixture design parameters make accurate prediction of concrete compressive strength challenging. In this study, TabPFN, a pre-trained foundation model for tabular data, was applied to predict the compressive strength of fly ash concrete and compared with tuned Random Forest, support vector regression, an artificial neural network, LightGBM, CatBoost, Ridge regression, and Abrams empirical regression. A dataset containing 1062 samples and eight mixture-level variables was used for model development and evaluation. Predictive performance was assessed using the coefficient of determination, mean absolute error, and root mean square error over 100 repeated random splits. The results showed that TabPFN achieved the best overall performance, with an average coefficient of determination of 0.9329, a mean absolute error of 3.2758 MPa, and a root mean square error of 4.6678 MPa. Compared with the strongest tuned gradient-boosting baseline, CatBoost, TabPFN reduced the mean absolute error and root mean square error by 0.8768 MPa and 0.8560 MPa, respectively. Furthermore, repeated-split conformal prediction demonstrated reliable uncertainty quantification, with an average prediction interval coverage probability of 0.9615 and a mean prediction interval width of 23.4554 MPa. SHAP analysis identified the water-to-cement ratio, mortar strength, and water-to-binder ratio as important variables, while additional multicollinearity and feature ablation analyses indicated that correlated ratio variables should be interpreted cautiously. The results indicate that TabPFN provides an accurate, robust, and uncertainty-aware framework for preliminary prediction of 28-day fly ash concrete compressive strength.
Zhihao Zhao, Jinjin Wang, Guohui Ma et al.· Buildings· 0 citations
Accurate prediction of the mechanical strength of Basalt Fiber Reinforced Concrete (BFRC) is critical for structural design, safety assessment, and the advancement of sustainable infrastructure in civil engineering. Traditional prediction methods often fail to capture the nonlinear relationships between BFRC mix proportions and resulting strength characteristics, leading to unreliable estimations. To address this limitation, this study proposes the Optimized Moment Balanced Machine (OMBM), an advanced machine learning model developed to improve the predictive accuracy of BFRC strength parameters. The model was trained and evaluated using key input features, including cement content, silica fume, fly ash, superplasticizer, water, aggregate composition, and fiber property parameters. The performance of the OMBM was benchmarked against four established machine learning models, such as Least Squares Support Vector Machine (LSSVM), Backpropagation Neural Network (BPNN), K-Nearest Neighbors (KNN), and Linear Regression (LR). Results from ten-fold cross-validation show that OMBM consistently outperforms the comparison models across five evaluation metrics. It achieved the lowest RMSE (2.411), MAE (1.788), and MAPE (4.08%), along with the highest values for correlation coefficient (R = 0.978), and coefficient of determination (R2 = 0.956). Furthermore, the OMBM achieved a Reference Index (RI) score of 1.000, which confirms its position as the leading predictive model within this comparative framework. These results confirm the robustness and reliability of the proposed OMBM model, making it a highly effective tool for accurate strength prediction of BFRC. This approach offers significant potential for the advancement of sustainable infrastructure by enabling more accurate and efficient use of concrete materials.
R. R. Khasani, Ferry Hermawan, Yuliana Usman· IOP Conference Series: Earth...· 0 citations
There has been a rise in the need of concrete leading to high consumption of cement and emission of carbon. Sustainable concrete incorporating supplementary cementitious materials (SCMs) is an alternative that is eco-friendly, but the mechanical behavior is complicated and hard to predict using the traditional tests. The work presents a framework involving machine learning because of predicting the mechanical properties of sustainable concrete, namely, compressive, split tensile, and flexural strength. Four models such as Linear Regression, Support Vector Regression, Random Forest, and Artificial Neural Network were constructed based on a data of nearly 1000 sustainable concrete mixes. To estimate the model performance, R 2, RMSE and MAE were used. The findings indicated that ANN and RF had the greatest prediction accuracy. The feature analysis established the most influential factors to be water content, cement dosage, replacement ratio of SCM, and curing age. The mix design approach proposed here is a fast, economical, and sustainable approach to the design of concrete mix.
Arti Chouksey, Santosh Reddy P, P. S et al.· 2026 International Conferenc...· 0 citations
ABSTRACT This study establishes four popular data-driven techniques – Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and Gradient Boosting (GB) for predicting the flexural strength (FS) of high-performance steel fiber-reinforced cementitious composites (FRCCs). A database containing 156 experimental records was used to train and test predictive models with nine feasible input variables. The results demonstrated that the GB model was the best predictor for estimating the FS of FRCCs. The GB model maintained satisfactory predictive accuracy after 10-fold cross-validation, achieving an average R2 of 0.834 on the validation folds generated from the training dataset. The predictive capability of GB model remained stable at 600 Monte Carlo simulations. The Shapley Additive Explanations (SHAP) method and partial dependence plots (PDP) indicated that fiber volume content was the most influential factor affecting FS predictions. To validate the accuracy of the GB model, a single-point case study was conducted on three specimens subjected to a three-point bending load. The discrepancy between the experimental values and GB predictions corresponded to an absolute prediction error of 1.21%, highlighting the accuracy of the developed model. Finally, a cloud-based web application was developed to provide a convenient tool for practical FS prediction of FRCCs.
Duy-Liem Nguyen, Tan-Duy Phan· Journal of Structural Integr...· 0 citations
A comparative framework evaluating nine regression algorithms using the UCI Concrete Compressive Strength dataset, jointly integrating correlation-corrected statistical validation, multi-model Bayesian optimization, and domain-informed feature engineering with SHAP interpretation, rarely combined in prior concrete-strength studies.
Musthafa 'Abduh Fakhruddin, Sri Winarno, Acun Kardianawati· IDEALIS : InDonEsiA journaL...· 0 citations
The ability to predict concrete compressive strength is important in early-stage mix design screening. Thus, the predictions made from these models must match the actual data that was used to train and validate them. Therefore, this study assesses machine learning models using the publicly available UCI concrete compressive strength data set that has 1030 tabular entries. The tabular entries are defined by the following variables; cement, blast furnace slag, fly ash, water, superplasticiser, coarse aggregate, fine aggregate, curing time and measured compressive strength. As such, the study is framed as a transparent tabular prediction benchmark rather than as an experimental evaluation of microsilica and rubber aggregate based concretes. It does not claim to have evaluated any new test pieces, nor does it make any claims regarding SEM imaging or microstructural measurement of those test pieces. Further, there is no claim related to the use of any durability testing procedures. Several machine learning algorithms including support vector regression (SVR), random forest, XGBoost, artificial neural network (ANN) and an optimised tabular ensemble (TE), were each developed under the same leakage-controlled validation methodology. Performance metrics were provided based on both the native test set(s) and a common subset of the test set(s). Metrics included R
2
, RMSE, MAE, MAPE along with residual diagnostic and graphical error analyses. The new framework also emphasizes reproducibility, fairness of comparison and transparency of data domain limitations. In addition to supporting computer-based screening of conventional concrete strength databases, its results indicate what will be required for future studies that contain micro silica, rubber aggregates, microstructural measurements and/or durability measurements in their respective databases.
V. Vairagade· Journal of Materials Science...· 0 citations