通过设计线性分类器；最小风险贝叶斯分类器；监督学习法分层聚类分析；K－L变换提取有效特征，设计支持向量机对给定样本进行有效分类并分析结果。-By designing a linear classifier　minimum risk Bayes classifier　supervised learning method hierarchical cluster analysis　K-L transform to extract efficient features, designed to support vector machines for effective classification of a given sample and analyze the results.