An Introduction to Recursive Partitioning Using the RPART Routines ...

summary(fit)
Step{VI. Plot a standard version of the plot with some basic information.
plot(fit)
text(fit,use.n=T)
Step{VII. Create a prettier version of the tree.
post(fit,file="")
3 Rpart model options
This section examines the various options that are available when tting the model
rpart. The options are listed below with a brief explanation, then explored further
with actual examples.
The central tting function is rpart, whose main arguments are
formula: the model formula, as in lm and other S model tting functions. The
right hand side may contain both continuous and categorical (factor) terms.
If the outcome y has more than two levels, then categorical predictors must
be t by exhaustive enumeration, which can take avery long time.
data, weights, subset: as for other S models.
method: the type of splitting rule to use. Options at this point are classication,
anova, Poisson, and exponential.
parms: a list of method specic optional parameters. Poisson splitting has a single
parameter, the coecient of variation of the prior distribution on the rates
(shrink). It is used to prevent problems if nodes end up with 0 events. Usually
not changed (default=1). For classication, the list can contain any of:
prior{ thevector of prior probabilities
loss{ the loss matrix
split{ the splitting index
The priors must be positive and sum to 1. The loss matrix must have zeros
on the diagonal and positive o-diagonal elements. The splitting index can be
`gini' or `information'.
4