The finite impulse response (FIR) neural network technique is applied as a
nonlinear predictor to extrapolate a time-series signal from FDTD simulatio
ns. The FIR neural networks are trained by the temporal backpropagation lea
rning algorithm. The model is applied to predict time-domain evolution in t
he FDTD simulation of a wave-guide filter. (C) 1999 John Wiley & Sons, Inc.