中文说明：采用cross validation的思想可以在某种意义下得到最优的参数,可以有效的避免过学习和欠学习状态的发生,最终对于测试集合的预测得到较理想的准确率.采用实例验证表明,用cross validation选取出的参数来训练SVM得到的模型比随机的选取参数训练SVM得到的模型在最后分类预测上更有效.
English Description：
Used cross validation of thought can in a species meaning Xia get optimal of parameter, can effective of avoid had learning and owes learning State of occurred, eventually for test collection of forecast get more ideal of accurate rate. used instance validation showed that, with cross validation selected out of parameter to training SVM get of model than random of selected parameter training SVM get of model in last classification forecast Shang more effective.