cvloss or kfoldloss for regression tree?

Hi,

I'm a bit confused with 'cvloss' and 'kfoldLoss'.

1. kfoldLoss

https://uk.mathworks.com/help/stats/regressionpartitionedmodel.kfoldloss.html

Syntax: L = kfoldLoss(cvmodel) returns the cross-validation loss of cvmodel.

load carsmall
>> XX = [Displacement Horsepower Weight];
>> YY = MPG;
>> cvmodel = fitrtree(XX,YY,'crossval','on');
>> L = kfoldLoss(cvmodel,'mode','average')

L =

   30.3578

Default: 'mse', mean square root.

2. cvloss

https://uk.mathworks.com/help/stats/regressiontree.cvloss.html

Syntax: E = cvloss(tree) returns the cross-validated regression error (loss) for a regression tree.

>> load carsmall
>> X = [Displacement Horsepower Weight];
>> Mdl = fitrtree(X,MPG);
>> rng(1);
>> E = cvloss(Mdl)

E =

   25.7383

First, both cases used same predictors and same response, why there is a difference between L and E outcomes?

Second, function 'fitrtree' by default 'crossval' is turned 'off'. In the 'cvloss' example, noticed that 'Mdl = fitrtree(X,MPG);' didn't turn 'crossval' on, how does it have anything to do with cross-validated regression? It is not even turned on.

Third, how are both kfoldLoss and cvloss calculated? looks that they both use MSE but giving completely different results.

Answer 1

[Jeremy Brecevic]

Unlike cvloss, kfoldLoss does not return SE,Nleaf, or BestLevel. kfoldLoss also does not allow you to examine any error other than the classification error.

Found here: https://www.mathworks.com/help/stats/classificationtree.cvloss.html under alternatives