cv.tree(object, rand, FUN = shrink.tree, ..., big = F)
tree.
This is assumed to be the result of some function that produces
an object with the same named
components as that returned by the
tree() function.
tree.
All elements of
rand are in set 1:K for K-fold cross-validation.
An entry of 5, say, in the first position indicates that the first
observation will be in the test set in the
5th cross-validated replicate and in the training set for all K-1
remaining replicates.
If
rand is missing, random 10-fold cross-validation is performed.
tree.sequence.
Currently,
prune.tree and
shrink.tree are the only
such functions available.
FUN, in particular the argument
k, the
cost-complexity parameter if
FUN is
prune.tree or the shrinkage
parameter if
FUN is
shrink.tree. See these functions for details
on the use of
k.
big is
TRUE,
object is attached in
frame = 1 before
deviance is calculated and detached before the value is returned.
tree.sequence is returned.
The object contains the following components:
A
plot method exists for objects of this class.
It displays the value of the deviance for each tree in the sequence.
An additional axis displays the values of the sequencing parameter
for each tree.
Chambers, J.M., and Hastie, T.J. (1991). Statistical Models in S, pp. 393--395 and 409--410
z.auto <- tree(Mileage ~ Weight, car.test.frame) zcv <- cv.tree(z.auto, k=(1:9)/10) plot(zcv)