Machine learning for predicting strength properties of waste iron slag concrete.

This study investigates the utilization of waste iron slag (WIS) as a sustainable alternative in concrete production to reduce environmental impact and preserve natural resources. The experimental investigation of WIS-incorporated concrete focused on compressive and tensile strength with machine learning (ML) models for prediction. Among the tested ML algorithms, Decision Tree (DT) and XGBoost showed the highest accuracy (R = 0.95135) in predicting concrete strength properties, while models like SVM and Symbolic Regression underperformed. Experimental results indicate that up to 20 % WIS replacement maintains adequate strength, whereas higher proportions reduce structural integrity. A ranking score index and cost analysis confirmed the technical and economic feasibility of using WIS in concrete. Cost analysis demonstrated substantial cost savings with 25 % WIS incorporation, confirming its economic feasibility. Integrating experimental data with ML predictions highlights WIS's potential for sustainable concrete applications, enabling optimized mix designs and reduced reliance on physical testing. Future work should address limitations, including dataset expansion and the exploration of additional durability and mechanical properties to validate WIS's practicality in construction further.