How can I compute statistical significance with cross-validation?

Hello

I've done nested cross-validation with 10 outer runs and 10 inner runs to evaluate different classifiers. In the end I average the accuracies of the 10 outer cross-validation runs to get a final accuracy for one classifier.

Now, I would like to do some statistical test among the classifiers to see if one classifier performs better than another.

How can this be done in Matlab (or in general) based on the cross-validation accuracies?

[Esmeralda Ruiz Pujadas]

Hello, I am having the same issue. It seems no solution. According to a paper, the problem of nested cross-validation, you cannot apply statistical measurements afterwards because you have different parameters in each iteration. Were you able to implement an alternative?.