reg_rf: Robust random forests.

In rhli-Hannah/RCATE: Robust Estimation of Heterogeneous Treatment Effect

Description

reg_rf fit robust random forests based on weighted median splitting rule.

Usage

reg_rf(formula, n_trees = 50, feature_frac = 1/2, data, weights, minnodes = 5)

Arguments

formula	an object of class "formula".
n_trees	number of trees. Default is 50.
feature_frac	fraction of features used in each split. Default is 1/2.
data	a training data frame.
weights	an optional vector of weights.
minnodes	minimal leaf size of tree. Default is 5.

Value

a list of components

• fit - estimation method.

• importance - vector of variable importance.

• tree - tree structure and needed parameters.

• data - training data.

• nodepreds - leaf node predictions.

rhli-Hannah/RCATE documentation built on Aug. 26, 2020, 9:40 a.m.