采用svm来做分类，一般能得到较满意的结果，但用svm做分类预测时需要调节相关的参数才能得到比较理想的预测分类准确率，那么svm的参数该如何选取？该程序主要说明如何更好地提升分类器性能。(Use svm to do the classification, the general can get more satisfactory results, but when using svm to do classification prediction need to adjust the relevant parameters in order to get the ideal prediction classification accuracy, then svm parameters how to choose? The program mainly shows how to improve the performance of classifier.)