Discussion of: “Artificial neural network and mathematical modeling comparative analysis of nonisothermal diffusion of moisture in wood” by Stavros Avramidis and Hongwei Wu

摘要

Aspects of data processing are discussed with particular emphasis on data division into sets for network training and network testing. The close relationship between the data used in artificial neural network (ANN) development and model applicability is pointed out. Based on the data also used in the original paper a neural network is developed and ANN testing output is discussed and compared with the results described in the original paper. Conclusions are drawn with respect to the applicability of the model.