Support Vector Machines based Modelling of Concrete Strength

machines in predicting the compressive strength of high strength concrete. The method is applied to predict the 28-days compressive strength of 181 high strength concrete data from the work already reported in literature. Radial basis function (RBF) and polynomial kernels are used with support vector machines. For this data set, RBF kernel works well in comparison to polynomial kernel based support vector machines and provide a root mean square error of 2.31 (correlation coefficient = 0.996) for 28-days strength prediction. Another high strength concrete data set (190 data) obtained by experimental investigation in the laboratory, are also used to judge the performance of support vector machines and provided a correlation coefficient of 0.994 (root mean square error = 0.71) that suggests that RBF kernel based support vector machine works well for this data set data set also. Results with both data sets suggest that support vector machine based modeling approach can effectively be used in predicting the compressive strength of high performance concrete.