载入数据train_x,train_y,test_x,test_y后,用下面两条命令做回归预测model = svmtrain(train_y,train_x,'-s 3 -t 2 -c 15000 -g 128 -p 0.001' );[predict_y,pmse,dec_value] = svmpredict(test_y,test_x,model);结果发现预测值predict_y里所有值都一样,平方相关系数也有问题Squared correlation coefficient = 1.#INF (regression)。后来加了数据的归一化[train_x,test_x,ps] = scaleForSVM(train_x,test_x,0,1) [train_y,test_y,ps] = scaleForSVM(train_y,test_y,0,1)平方相关系数算出值了,但是predict_y里仍然所有值都一样?!直接用faruto封装好的[predict_Y,mse,r] = SVR(train_y,train_x,test_y,test_x,Method_option);则没有上述问题。大伙给支支招啊!