Cement-based materials, particularly sandcrete, play a crucial role in the construction industry, where demand for sustainable and high-performance materials is increasing. Fly ash, a byproduct of coal combustion, has gained attention as a supplementary cementitious material (SCM) to improve the sustainability of these materials. However, predicting the compressive strength of fly ash-blended mortars, which is essential for ensuring the structural integrity and durability of construction materials, remains a challenge. This study aims to address this gap by leveraging advanced machine learning techniques to predict the compressive strength of fly ash blended sandcrete, using key input variables such as aggregate-to-binder ratio (Agg/B), fly ash-to-binder ratio (FA/B), water-to-binder ratio (W/B), and curing time. While previous research has focused on conventional cement mortars, few studies have integrated these key variables using machine learning models. The goal of this research is to develop a reliable and accurate predictive model for compressive strength, filling a gap in existing literature regarding the influence of multiple input variables on mortar performance. The study employs four machine learning models - Artificial Neural Networks (ANN), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), and Extreme Gradient Boosting (XGB) - to evaluate their performance in predicting compressive strength. The results show that XGB outperforms other models, achieving an R² of 0.84 for training data and 0.74 for testing data, along with the lowest prediction errors. These findings demonstrate the potential of machine learning models, particularly XGB, in optimizing mix designs and improving sustainability in construction, offering valuable insights for future material innovations.

fly ash, sandcrete, compressive strength, machine learning, extreme gradient boosting