How do you obtain a vector of predicted classes generated after cross-valiation of a decision tree?

I have created a decision tree (T) to predict classes, Y, based on features X. I would like to cross-validate the decision tree. When I do so, I can only get the overall cross-validation error, not the vector of classes as they are predicted from the cross-validation.

Here is the code I am trying to use: T3 = classregtree(X,Y,'method','classification'; view(T3); cp = cvpartition(Y,'k',10);

dtClassFun = @(xtrain,ytrain,xtest)(eval(classregtree(xtrain,ytrain),xtest));

crossvalerrorrT = crossval('mcr',X,Y,'predfun',dtClassFun,'partition',cp)

This will output a scalar overall cross-validation error.

How can I output a vector of cross-validation predicted classes (so that I can look at cross-validation error for each class separately)?

Answer 1

[Tom Lane]

Consider this:

load fisheriris
cp = cvpartition(species,'k',10);
F = @(xtr,ytr,xtest,ytest){ytest, eval(classregtree(xtr,ytr),xtest)};
result = crossval(F,meas,species,'partition',cp)

This uses the other syntax of crossval. The function F packages up the observed and fitted Y values for each fold of the 1-fold cross-validation. The results is a cell array with ten rows, one for each fold, with each row containing the corresponding observed and fitted values.

Alternatively you could write the cross-validation loop yourself. Loop over j and get test(cp,j) and training(cp,j) each time.

[Kelly]

Very helpful, thank you!

Just to clarify one point:

Column 1 in 'result' is the observed y, and column 2 in 'result' is the predicted/fitted y for each fold cross-validation?

[Tom Lane]

That's right.

Answer 2

[Ilya]

If you have 11a or later, you can use ClassificationTree:

load fisheriris;
cvtree = ClassificationTree(meas,species,'kfold',10);
Yfit = kfoldPredict(cvtree);