作者: Murat Simsek , N. Serap Sengor
DOI: 10.1002/JNM.656
关键词:
摘要: SUMMARY This paper presents two new methods, space mapping (SM) with prior knowledge input (PKI-D) difference and compound mapping-based neuromodeling. Both methods combine powerful techniques, neuromodeling PKI-D difference. The knowledge-based modeling in the RF/microwave literature merge about device to be modeled neural network structures while a method, SP, focuses on reducing computational burden. main advantage of proposed over these already existing are their better extrapolation capability reduced number training set data. simulation results obtained reveal that both decrease cost improve output performance SP-based Copyright # 2007 John Wiley & Sons, Ltd.