Predicting the compressive strength of fly ash concrete using the XGBoost machine learning model

Abstract

Fly ash concrete is a type of construction material in which a portion of Portland cement is replaced with fly ash—a byproduct collected from coal combustion in thermal power plants. Using fly ash at an appropriate ratio contributes to a more sustainable concrete mix and offers significant environmental benefits. In this study, the XGBoost machine learning model is applied to predict the compressive strength of fly ash concrete, with the aim of optimizing the mix proportions and ensuring practicalperformance and quality. The model is trained on a dataset with seven input parameters: cement, sand, coarse aggregate, water, fly ash, admixture, and curing time; while the compressive strength is used as the output target. The model’s performance is evaluated using metrics such as mean absolute error (MAE), root mean squared error (RMSE), and the coefficient of determination (R²). The results show that the XGBoost model is an effective approach for predicting the compressive strength of fly ash concrete,achieving training results of (R² = 0,896; RMSE = 4,213; MAE = 3,133) and validation results of (R² = 0,868; RMSE = 4,892; MAE = 3,401).