Application of SVM and BCG-FFT Method for the Parameter Reconstruction of Composite Conducting-Dielectric Cylinder

作者：Zhang, Q. H.^*; Chen, H. T.

来源：Applied Computational Electromagnetics Society Journal, 2014, 29(5): 391-399.

• Biconjugate Gradient Fast Fourier Transform (BCG-FFT)
• composite conducting-dielectric cylinder
• parameter reconstruction and Support Vector Machine (SVM)

In this paper, Support Vector Machine (SVM) technique is used to reconstruct the geometric and dielectric characteristics of composite conducting-dielectric cylinder. To this aim, the scattered electric fields at a number of observation points by composite conducting-dielectric object under the different object parameters are calculated by stabilized Biconjugate Gradient Fast Fourier Transform method (BCG-FFT) and provide to SVM as input training samples, while the output of the SVM are the characteristics of the objects. In numerical results, the proposed technique is applied successfully to the reconstruction of the geometric and dielectric parameters of composite conducting-dielectric cylinder. The effectiveness of the SVM method is evaluated and also in comparison with the Neural Network (NN) based approaches.

• 出版日期2014-5
• 单位三峡大学