Predictive modeling of crumb rubber-modified mortar: insights from ANN, LR, RF, and M5P methods

Abstract

This study investigates the feasibility of using crumb rubber as a partial sand replacement in cement mortar, aiming to address environmental challenges associated with tire waste while contributing to sustainable construction practices. The experimental phase involved preparing cement mortar samples with varying percentages of crumb rubber and analyzing the resulting compressive strength. Crumb rubber substitution at levels of up to 7.5% in a 1:5 to 1:6 mix proportion resulted in practical compressive strengths between 2–6 MPa, suitable for certain applications in construction. The compressive strength reduction associated with increased crumb rubber was offset by improved durability characteristics, including enhanced ductility and energy absorption. To model and predict compressive strength effectively, four machine learning approaches—Artificial Neural Network (ANN), Random Forest (RF), Linear Regression (LR), and M5P tree—were implemented. The ANN model emerged as the most effective with respect to testing data, with performance metrics including Coefficient of Correlation (CC) values of 0.9998, Nash–Sutcliffe Efficiency (NSE) values 0.9959, least Root Mean Squared Error (RMSE) of 0.2125, least Scattering Index (SI) of 0.041 and least Mean Absolute Error (MAE) of 0.1693. Sensitivity analysis further highlighted crumb rubber percentage as a critical factor influencing compressive strength, underscoring the potential for targeted optimization. The findings suggest that incorporating crumb rubber in mortar can balance sustainability goals with material performance, especially when paired with advanced predictive modeling. Future work is recommended to optimize formulations by varying water-cement ratios or introducing plasticizers to enhance the strength of rubber-modified mortar. This research highlights a promising pathway for reusing waste materials in construction, contributing to both environmental and structural engineering fields.