rpartXse {DMwR2} | R Documentation |
Obtain a tree-based model
Description
This function is based on the tree-based framework provided by the
rpart
package (Therneau et. al. 2010). It basically, integrates
the tree growth and tree post-pruning in a single function call. The
post-pruning phase is essentially the 1-SE rule described in the CART
book (Breiman et. al. 1984).
Usage
rpartXse(form, data, se = 1, cp = 0, minsplit = 6, verbose = F, ...)
Arguments
form |
A formula describing the prediction problem |
data |
A data frame containg the training data to be used to obtain the tree-based model |
se |
A value with the number of standard errors to use in the post-pruning of the tree using the SE rule (defaults to 1) |
cp |
A value that controls the stopping criteria used to stop the initial tree growth (defaults to 0) |
minsplit |
A value that controls the stopping criteria used to stop the initial tree growth (defaults to 6) |
verbose |
The level of verbosity of the function (defaults to F) |
... |
Any other arguments that are passed to the |
Details
The x-SE rule for tree post-pruning is based on the cross-validation estimates of the error of the sub-trees of the initially grown tree, together with the standard errors of these estimates. These values are used to select the final tree model. Namely, the selected tree is the smallest tree with estimated error less than the B+x*SE, where B is the lowest estimate of error and SE is the standard error of this B estimate.
Value
A rpart
object
Author(s)
Luis Torgo ltorgo@dcc.fc.up.pt
References
Therneau, T. M. and Atkinson, B.; port by Brian Ripley. (2010). rpart: Recursive Partitioning. R package version 3.1-46.
Breiman, L., Friedman, J., Olshen, R., and Stone, C. (1984). Classification and regression trees. Statistics/Probability Series. Wadsworth & Brooks/Cole Advanced Books & Software.
Torgo, L. (2016) Data Mining using R: learning with case studies, second edition, Chapman & Hall/CRC (ISBN-13: 978-1482234893).
See Also
Examples
data(iris)
tree <- rpartXse(Species ~ ., iris)
tree
## A visual representation of the classification tree
## Not run:
prettyTree(tree)
## End(Not run)