Research on Improved FSV Method Based on Wavelet Function Transform

Feature Selective Validation method is the core algorithm of model verification and simulation verification established by IEEE Standard 1597.1, in order to quantitatively evaluate the reliability of electromagnetic simulation results. Aiming at the failure problem of data with transient component in Feature Selective Validation method, this paper systematically analyzes the negative effect of Fourier transform in this process, and proposes an improved Feature Selective Validation method based on DB wavelet function transform. Based on the immune algorithm, the Feature Difference Measure coefficient of the improved Feature Selective Validation method is corrected to ensure its consistency with the traditional feature selection verification method in the evaluation of 8 basic problems. Finally, according to 7 sets of evaluation questions involving transient components from the Polytechnic University of Catalonia, it is verified that the proposed method is closer to the evaluation results of experts with professional electromagnetic simulation background.