Information Technology Journal1812-56381812-5646Asian Network for Scientific Information10.3923/itj.2011.1999.2003MaCheng SanYe 1020111010In order to widen the range of processing signals for fuzzy neural network, a kind of fuzzy process neural network based on orthogonal basis function is proposed. By inducting the orthogonal basis function into input space, the input function can be feature expanded. Meanwhile the weight vectors are also expanded under the same orthogonal basis function. The operation process of space aggregation and time cumulative can be simplified by using the orthogonality of basis function. The back propagation algorithm is used as learning algorithm. Simulation results show the great approximation ability of the fuzzy process neural network. And further experiments show that the network is sensitive to the number of fuzzy rules. The best approximation accuracy can be obtained only by choosing the proper values, such as k = 4 in this study.]]>Amjady, N.,2006Pindoriya, N.M., S.N. Singh and S.K. Singh,2008Lin, F.J. and P.H. Chou,2009Lin, C.T., W.C. Cheng and S.F. Liang,2005Han, H. and J. Qiao,2010Coyle, D., G. Prasad and T.M. McGinnity,2009Roh, S.B., W. Pedrycz and S.K. Oh,2007He, X.G. and J.Z. Liang,2000He, X.G. and S.H. Xu,2003Gang, D., Z. Shi-Sheng and L. Yang,2008