中文摘要：

介绍了一种基于Fisher线性判别的非线性分类方法——核Fisher，其主要思想是首先把样本映射到特征空间F，然后在此空间进行Fisher线性判别，隐含地实现了原输入空间的非线性判别；同时利用SVM对分类阚值进行估计，实现了对两类样本最大程度的区分。通过仿真可以得出这一判别方式有利于确定两类平均丢包率的区分阈值。

英文摘要：

Based on the Fisher linearity differentiation, a kind of non-linearity sorting algorithm is introduced. The algorithm maps the swatch to a character space, and then utilizes Fisher linearity differentiation to realize the nonlinearity differentiation in the original input space. And the sorting threshold value is estimated using support vector machine（SVM）. Finally the most extent differentiation about the two kind swatches is achieved. Simulation results show that the algorithm can work well on differentiating the threshold value of two kinds of average packet loss ratios.