Fuzzy neural network modeling of material properties

Abstract In this paper, an adaptive fuzzy neural network has been developed for extracting fuzzy rules directly from experimental data for material property modeling. As an application example, a model has been developed to predict strength based on compositions and microstructure by the use of the adaptive neuro-fuzzy inference method. Compared with the results obtained by multiple statistical analysis, fuzzy regression and the generalized regression neural network, the fuzzy neural network shows good learning precision and generalization.