WSEAS Transactions on Information Science and Applications
Print ISSN: 1790-0832, E-ISSN: 2224-3402
Volume 22, 2025
Prediction of Concrete Compressive Strength with Artificial Intelligence Applications
Authors: , , ,
Search Articles
Abstract: In this study, we aimed to predict the compressive strength of concrete using artificial intelligence, specifically the CatBoost Regressor. To make sense of the model’s predictions, we used SHAP and LIME, which helped reveal how each factor influenced the results. We also applied a greedy feature selection approach to focus on the most important variables, improving both accuracy and computational efficiency. The dataset included key ingredients like cement, water, sand, and aggregates, along with the concrete’s age. Before running the model, we cleaned the data, filled in missing values, and normalized the features. Finally, we evaluated the model’s performance using metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and $$R^{2}$$, which together showed that the model was effective at capturing the underlying patterns.
Pages: 740-747
DOI: 10.37394/23209.2025.22.61