bartMachine (version 1.2.6)

rmse_by_num_trees: Assess the Out-of-sample RMSE by Number of Trees

Description

Assess out-of-sample RMSE of a BART model for varying numbers of trees in the sum-of-trees model.

Usage

rmse_by_num_trees(bart_machine, tree_list = c(5, seq(10, 50, 10), 100, 150, 200),
in_sample = FALSE, plot = TRUE, holdout_pctg = 0.3, num_replicates = 4, ...)

Value

Invisibly, returns the out-of-sample average RMSEs for each tree size.

Arguments

bart_machine

An object of class ``bartMachine''.

tree_list

List of sizes for the sum-of-trees models.

in_sample

If TRUE, the RMSE is computed on in-sample data rather than an out-of-sample holdout.

plot

If TRUE, a plot of the RMSE by the number of trees in the ensemble is created.

holdout_pctg

Percentage of the data to be treated as an out-of-sample holdout.

num_replicates

Number of replicates to average the results over. Each replicate uses a randomly sampled holdout of the data, (which could have overlap).

...

Other arguments to be passed to the plot function.

Author

Adam Kapelner and Justin Bleich

Examples

if (FALSE) {
#generate Friedman data
set.seed(11)
n  = 200 
p = 10
X = data.frame(matrix(runif(n * p), ncol = p))
y = 10 * sin(pi* X[ ,1] * X[,2]) +20 * (X[,3] -.5)^2 + 10 * X[ ,4] + 5 * X[,5] + rnorm(n)

##build BART regression model
bart_machine = bartMachine(X, y, num_trees = 20)

#explore RMSE by number of trees
rmse_by_num_trees(bart_machine)
}