forestr: Classification and Regression with Random Forest and Choice...

In andeek/forestr: Random forests with a user created splitting criterion

Description

Runs a random forest with user specified splitting criteria using the forestr function.

Usage

forestr(formula, data, mvars, B = 500, min_size, method, ...)

## S3 method for class 'forestr'
print(x, ...)

Arguments

formula, x	formula specification of forest variables, (for the print method, a forestr object)
data	dataframe with y and x variables
mvars	number of variables to sample in each node of the tree during building the random forest
B	number of bootstrap samples
min_size	minimum size of a terminal node
method	optional splitting method, currently "extremes" and "purity" implemented.
...	extra parameters to pass to rpart

Value

An object of class forestr with components

call	the original call to forest
type	one of regression or classification
predicted	the predicted values of the input data based on out of bag samples
importance	mean decrease in accurace over all classes for each variable in the model
votes	matrix giving the votes for each class on each observation (classification)
oob	out of bag error for the model including confusion matrix (classification)
trees	random forest tree objects

Examples

forestr(factor(cyl) ~ ., data = mtcars)

andeek/forestr documentation built on May 12, 2019, 2:40 a.m.