A statistic comparison between radial basis function neural network and backpropagation neural network in die casting process modeling

Nowadays, given its characteristics, the casting process is one of the most important manufacturing processes in the automotive industry. However, with the advancement of new technologies arises the need to apply tools to analyze the behavior of the process. The Radial Basis Function Neural Network (RBFNN) and the Back Propagation Neural Network (BPNN) are alternative tools to model manufacturing processes due to their good performance in prediction problems. The main aim of the present paper is to model the casting process through RBFNN and BPNN propose a model selection criterion. Both models turned out to be very similar in terms of $\mathrm{R}^{2}$, therefore, the AIC and the BIC criteria were applied to determine which neural network is the best model for the manufacturing process. It was found that the BPNN better represents the behavior of the process, however, a likelihood value was found that might not be the maximum.