The determination of kernel function parameter initial value of the KPCA based on adaptive contribution rate
[摘要] In this article, we start from the samples and the kernel principal component contribution rate adaptive point of view. Our aim is to research that how to choose the initial parameters of the mixed kernel function in comprehensive evaluation of KPCA. To some extent, it will overcome the shortcomings that the kernel function parameter is difficult to determine, and it provides an effective method that KPCA applies to determine the kernel function initial parameter of the multi-indexes comprehensive evaluation. We believe that it will be more widely applied to feature extraction with its advantages appear in specific issues, and makes the evaluation method more and more scientific and effective.
[发布日期] [发布机构] College of Science, North China University of Science and Technology, Tangshan; 063210, China^1
[效力级别] 生态环境科学 [学科分类]
[关键词] Comprehensive evaluation;Contribution rate;Initial parameter;Kernel function;Kernel function parameters;Kernel principal component;Mixed kernel function [时效性]
