Discussion on: "Prediction of mechanical properties of recycled aggregate concretes containing silica fume using artificial neural networks and fuzzy logic" [Comp. Mater. Sci. 42 (2008) 74]

A review on an experiment related to the mechanical properties of recycled aggregate concretes containing silica fume using artificial neural networks (ANN) and fuzzy logic, is presented. Artificial neural networks (ANN) and fuzzy logic (FL) models were used to predict compressive and splitting tensile strengths of recycled aggregate concretes containing silica fume. The experiment failed to consider basic conceptions of the FL principle. The model developed in the experiment has two neurons in the output layer, which are compressive strength and splitting tensile strength (STS), but it was mentioned by the researchers that ANN model has eight neurons in the input layer and three neurons in the output layer. The experiment also failed to discuss the issue of sensitivity to the initial weight assignments of back propagation learning algorithm of ANN and local minima issue.