Ứng dụng mô hình rừng cây ngẫu nhiên để dự đoán cường độ chịu nén của bê tông

Abstract

Cement concrete is the most widely used material today in construction projects from infrastructure to civil construction thanks to its good compressive strength and competitive price. Conducting experiments to determine compressive strength of cement concrete requires high costs and long implementation time. Therefore, the application of the random forest model is a branch of artificial intelligence to determine the compressive strength of cement concrete is very meaningful. Random forest model has been applied to train and test 1030 samples of compressive strength of cement concrete. The predicted results of the random forest model give relatively high accuracy in two training and testing cases with correlation coefficients R respectively 0.99 and 0.95. Therefore, it is feasible to apply artificial intelligence, particularly random forest models, to determine the compressive strength of cement concrete using many additives. The random forest model can also identify the most important factors affecting the compressive strength of cement concrete such as cement and age which is most affecting the compressive strength of cement concrete. The fly ash is the least affected factor on the compressive strength of cement concrete.