Bài báo khoa học

The concrete grade is designed based on the function of the concrete structure and its use conditions. However, accurate estimation of concrete grade is a challenging task because of the nonlinear relationship between constituent materials. This paper investigates the performance of several machine learning (ML) methods, including Naïve Bayes, K-Nearest Neighbors, Decision Tree, Random Forest, and XGBoost, for classifying the concrete grades. A database of concrete available in the literature is used to develop ML models. The results of the ML models are evaluated and compared to choose the best ML model for classifying the concrete grades. The XGBoost model outperforms other models with 100% and 92% accuracy for the training and test set, respectively. As a result, the developed XGBoost model can save time and cost to classify the concrete grades without conducting any concrete samples.