rsq.itree: Plots the Approximate R-Square for the Different Splits

In itree: Tools for classification and regression trees, with an emphasis on interpretability.

Description

Produces 2 plots. The first plots the r-square (apparent and apparent - from cross-validation) versus the number of splits. The second plots the Relative Error(cross-validation) +/- 1-SE from cross-validation versus the number of splits. Same as rsq.rpart but does some checking to make sure warnings/error is printed when the user attempts to call the function in cases where either cptable=NULL or does not have the correct meaning to make the plots useful. Identical to the rpart function.

Usage

rsq.itree(x)

Arguments

x	fitted model object of class itree. This is assumed to be the result of some function that produces an object with the same named components as that returned by the itree function.

Side Effects

Two plots are produced.

Note

The labels are only appropriate for the "anova" method. Further the cptable from which the r-squared values are taken are not appropriate if a penalty has been used. If this is the case, the method stops.

Examples

#rpart's example:
z.auto <- itree(Mileage ~ Weight, car.test.frame)
rsq.itree(z.auto)

Example output

itree is based on the code of rpart.
Bug reports should be directed to this package's maintainer, not rparts'.


Regression tree:
itree(formula = Mileage ~ Weight, data = car.test.frame)

Variables actually used in tree construction:
[1] Weight

Root node error: 1354.6/60 = 22.576

n= 60 

        CP nsplit rel error  xerror     xstd
1 0.595349      0   1.00000 1.02697 0.178133
2 0.134528      1   0.40465 0.52391 0.084021
3 0.012828      2   0.27012 0.40198 0.075862
4 0.010000      3   0.25729 0.41299 0.075400
Warning message:
In printcp(x) :
  cp is impacted by using penalties and is NOT comparable to unpenalized cp's.

itree documentation built on May 2, 2019, 7:25 a.m.