Composite conducting-dielectric cylinder parameters reconstruction by means of SVM

The parameters of composite conducting-dielectric 2-domain cylinder are reconstructed by means of support vector machine(SVM)and stabilized biconjugate gradient fast Fourier transform(BCG-FFT).The scattered electric fields by composite object is measured by using BCG-FFT,and then be fed into the SVM,after proper training,the geometric and electromagnetic parameters of composite cylinder are reconstructed in real-time.In the meantime,at the same conditions,the composite object parameters also are reconstructed by means of artificial neural networks(ANN)method.Finally,the difference of trained examples information influence on the reconstructed precision is compared and analyzed.Through comparison with the ANN method,the support vector machine method is valid for the composite object parameters reconstruction,and has higher precision.