Abstract: The advantage of a kernel method often depends critically on a proper choice of the kernel function.A promising approach is to learn the kernel from data automatically.In this paper，a novel method is proposed for learning the kernel matrix based on maximizing a class separability criterion that is similar to those used by Linear Discriminant Analysis（LDA） and Kernel Fisher Discriminant（KFD）.This paper proposes this approach when FSVM is used for multiclassification task.And the results of experiments on face data set shows that the method is feasible.

Key words: kernel learning, Support Vector Machine（SVM）, Fuzzy Support Vector Machine（FSVM）, Fisher discriminant criterion

摘要： 一个有效的核方法通常取决于选择一个合适的核函数。目前研究核方法的热点是从数据中自动地进行核学习。提出基于最优分类标准的核学习方法，这个标准类似于线性鉴别分析和核Fisher判别式。并把此算法应用于模糊支持向量机多类分类器设计上，在ORL人脸数据集和Iris数据集上的实验验证了该算法的可行性。

关键词: 核学习, 支持向量机, 模糊支持向量机, Fisher判别准则