performance evaluation of rbf networks with various variables to forecast the properties of sccs

In the present study, radial basis function (rbf) neural networks are applied to forecast the compressive strength and elastic modulus of self-compacting concrete (scc). to construct the models, different experimental specimens of diverse kinds of scc are gathered from the literature. the data used in the networks are classified into two different sets of input parameters. the results revealed that the proposed rbf models can accurately forecast the properties of sccs with low test error. furthermore, a comparison between models with two different sets of inputs proves that the selected parameters as input variables, straightly impress the precision of the networks, in the prediction of the intended outputs.