Predicting Compressive Strength of Concrete Incorporating Fly Ash, Blast Furnace Slag, and Superplasticizer Using Machine Learning Techniques

Concrete is the second most essential element in the construction industry, and its strength requirements vary based on the specific conditions of each project. However, determining the compressive strength of concrete involves laboratory tests, which wastes a lot of time and money. Researchers have developed machine learning models that predict the compressive strength of cement-based concrete having various mixes. In this research, the compressive strength of concrete incorporating fly ash, blast furnace slag, and superplasticizer is predicted using different machine learning models, namely, Linear Regression, Random Forest Regression, Decision Tree Regression, Extreme Gradient Boosting, Light Gradient Boosting, AdaBoost, and CatBoost Regression. Then the model’s performance is evaluated based on prediction accuracy and prediction error rates i.e. R square, MSE, RMSE, MAE, MSLE, and RMSLE. Comparing all seven models, CatBoost and XGBoost have shown the highest prediction accuracy (𝑅2=0.8948,𝑀𝐴𝐸=4.034,𝑅𝑀𝑆𝐸= 5.766 and 𝑅2=0.890,𝑀𝐴𝐸=3.978,𝑅𝑀𝑆𝐸=5.881) respectively. In conclusion, the CatBoost and XGBoost are the best machine learning models for predicting the compressive strength of concrete with the highest prediction accuracy, and lowest modeling error.