【摘要】 如何选择最优或接近最优的核函数使分类错误率降低,是KPCA(Kenel Principle Com-ponent Analysis)应用于特征提取的关键。本文在研究了文化算法(Cultural Algorithms,CA)相关文献的基础上,提出了一种训练核函数参数的文化算法流程,实现了KPCA和CA的集成,有效地提高了核函数的优化选择。仿真结果表明该方法具有较好的结果和更少的计算量。更多还原

【Abstract】 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.更多还原