Recursive Partitioning and Regression Trees
rpart(formula, data, weights, subset, na.action = na.rpart, method, model = FALSE, x = FALSE, y = TRUE, parms, control, cost, ...)
- a formula, with a response but no interaction
terms. If this a a data frome, that is taken as the model frame
- an optional data frame in which to interpret the variables named in the formula.
- optional case weights.
- optional expression saying that only a subset of the rows of the data should be used in the fit.
- the default action deletes all observations for which
yis missing, but keeps those in which one or more predictors are missing.
- one of
methodis missing then the routine tries to make an intelligent guess. If
yis a survival object, then
method = "exp"is assumed, if
yhas 2 columns then
method = "poisson"is assumed, if
yis a factor then
method = "class"is assumed, otherwise
method = "anova"is assumed. It is wisest to specify the method directly, especially as more criteria may added to the function in future.
methodcan be a list of functions named
eval. Examples are given in the file tests/usersplits.R in the sources, and in the vignettes User Written Split Functions.
- if logical: keep a copy of the model frame in the result?
If the input value for
modelis a model frame (likely from an earlier call to the
rpartfunction), then this frame is used rather than constructing new data.
- keep a copy of the
xmatrix in the result.
- keep a copy of the dependent variable in the result. If
modelis supplied this defaults to
- optional parameters for the splitting function.
Anova splitting has no parameters.
Poisson splitting has a single parameter, the coefficient of variation of
the prior distribution on the rates. The default value is 1.
Exponential splitting has the same parameter as Poisson.
For classification splitting, the list can contain any of:
the vector of prior probabilities (component
prior), the loss matrix (component
loss) or the splitting index (component
split). The priors must be positive and sum to 1. The loss matrix must have zeros on the diagonal and positive off-diagonal elements. The splitting index can be
information. The default priors are proportional to the data counts, the losses default to 1, and the split defaults to
- a list of options that control details of the
- a vector of non-negative costs, one for each variable in the model. Defaults to one for all variables. These are scalings to be applied when considering splits, so the improvement on splitting on a variable is divided by its cost in deciding which split to choose.
- arguments to
rpart.controlmay also be specified in the call to
rpart. They are checked against the list of valid arguments.
This differs from the
tree function in S mainly in its handling
of surrogate variables. In most details it follows Breiman
et. al (1984) quite closely. R package tree provides a
An object of class
Breiman L., Friedman J. H., Olshen R. A., and Stone, C. J. (1984) Classification and Regression Trees. Wadsworth.
fit <- rpart(Kyphosis ~ Age + Number + Start, data = kyphosis) fit2 <- rpart(Kyphosis ~ Age + Number + Start, data = kyphosis, parms = list(prior = c(.65,.35), split = "information")) fit3 <- rpart(Kyphosis ~ Age + Number + Start, data = kyphosis, control = rpart.control(cp = 0.05)) par(mfrow = c(1,2), xpd = NA) # otherwise on some devices the text is clipped plot(fit) text(fit, use.n = TRUE) plot(fit2) text(fit2, use.n = TRUE)
# Set random seed. Don't remove this line. set.seed(1) # Take a look at the iris dataset str(iris) summary(iris) # A decision tree model has been built for you tree <- rpart(Species ~ Sepal.Length + Sepal.Width + Petal.Length + Petal.Width, data = iris, method = "class") # A dataframe containing unseen observations unseen <- data.frame(Sepal.Length = c(5.3, 7.2), Sepal.Width = c(2.9, 3.9), Petal.Length = c(1.7, 5.4), Petal.Width = c(0.8, 2.3)) # Predict the label of the unseen observations. Print out the result. predict(tree, unseen, type = "class")