Abstract:  In order to compare the approximation performances of GRNN and RBFNN to binary functions, GRNN and RBFNN are first established through computer programming in this paper. A specific binary function is taken as an example to be approximated using the above two neural networks respectively. The simulation results show that in approximation to a binary function, compared with RBFNN, GRNN has higher precision, faster convergence speed, and better approximation ability. Thus it provides a good method to solve the problem of binary nonlinear function approximation.

 Key words:  GRNN, RBFNN, binary function, approximation, simulation