KPCA Based on Cultural Algorithms Feature Extraction

How to choose the best or near kernel function to reduce classifications error rate is the key of KPCA applied to extract nonlinear feature components.In this paper,on the basis of research of CA,we propose a programmer flow of CA used for training kernel function and build CA-KPCA.This approach can effectively optimize kernel function.Simulation results show that produces highly competitive results at a relatively low computational cost.