Chiang Mai Journal of Science

This paper focused on the application of  design of  experiments to determine optimize parameters for multilayer-perceptron artificial neural network trained with back-propagation for modeling purpose.  Artificial neural networks (ANNs) for modeling have been widely used in various fields because of  its ability to ‘learn’ from examples. The accuracy of  ANN model depends very much on the setting of network parameters, such as number of neurons, number of  hidden layers and learning rate. Most literatures in this area suggested trial-anderror method for parameters setting which are time consuming and non economical, whereas the optimal setting cannot be guaranteed. Consequently, design of  experiment techniques is generally required to optimize various processes. In this paper, as a case study, it was used to find optimum setting of ANN trained to model ferromagnetic material data. Interested characteristic was finite-sized ferromagnetic Curie temperature obtained from Monte Carlo simulation on two dimensional Ising spins.  The results indicated that design of  experiments is a promising solution to the mentioned problem. The issues arising from this case were also discussed.