A new method to determine the elements of GaAs MESFET from measured S-parameters using neural networks

Recently neural networks have been introduced to the microwave area as fast and flexible tools to microwave modeling, simulation and optimization. In this paper, a radial basis transfer function, combined with microwave empirical or semi-analytical information, is proposed. The microwave knowledge complements the capability of learning and generalization of neural networks by providing additional information that may not be adequately represented in a limited set of training data. Such knowledge becomes even more valuable when the neural model is used to extrapolate beyond training data region.