Kernel nonparametric discriminant analysis

In this paper, a kernelized version of nonparametric discriminant analysis is proposed that we name KNDA. The main idea is to first map the original data into another high dimensional space, and then to perform nonparametric discriminant analysis in the high dimensional space. Nonparametric discriminant analysis can relax the Gaussian assumption required for the classical linear discriminant analysis, and Kernel trick can further improve the separation ability. A group of tests on several UCI standard benchmarks have been carried out that prove our proposed method is very promising.