Tóm tắt:

The impact of different rice husk ash contents (5, 10, 20%) on mortar strength is examined at different elevated temperatures (150, 300, 450, 750oC). Based on a 45 experimental result data set, three machine learning algorithms including the Artificial Neural Network (ANN), the Least Squares Support Vector Regression (LS-SVR) and the Multivariate Adaptive Regression Splines (MARS) have been used to model the functional relationship between the mixture components and the compressive strength. As a result, it is shown that LS-SVR consists in the most capable approach for modeling mortar strength with a good value of coefficient of determination (R2) = 0.80. Accordingly, this machine learning approach is potential to be used in RHA contained mix design by construction engineers.