说明：  pca做特征降维，然后进行特征空间随机分割构造多个svm分类器，并行处理，对样本进行分类，基于特征空间的svm多分类器
(using pca reduce feature dimension,split feature space and then randomly divided over svm classifier construction, parallel processing, the samples were classified, based on multi-feature space classification svm )