中文摘要：

针对分类过程中噪声影响问题,提出了基于模糊支持向量的判别分析法。利用基于马氏距离的模糊c均值算法,对每个训练样本点赋予不同的隶属度。将模糊支持向量机与Fisher判别法相结合,达到降维去噪的目的。实验结果表明：提出的分类方法的错分率明显低于其他的方法,而且能减轻噪声对分类的不良影响,从而证明算法的有效性。

英文摘要：

For the negative impact of noise on the classification, the discriminant analysis algorithm is proposed based on fuzzy support vectors. Fuzzy c-means clustering algorithm based on Mahalanobis distance is used to determine membership. Then the fuzzy support vector machine combined with Fisher discriminant, achieves dimension reduction and mitigate effect of noise. Experimental results show that the proposed method is lower error rate than other methods, and reduces the negative impact of noise in the classifier. The presented approach is proved to be effective.