求教关于交叉验证

wolfsong

高手能帮我看看crossvalind函数的和help中给的例子什么意思吗？本人对这方面比较陌生
% Create a 10-fold cross-validation to compute classification error.
       load fisheriris
       indices = crossvalind('Kfold',species,10);
       cp = classperf(species);
       for i = 1:10
           test = (indices == i); train = ~test;
           class = classify(meas(test,:),meas(train,:),species(train,:));
           classperf(cp,class,test)
       end

哪为高手对交叉验证有研究吗？急用！