grf (version 0.10.0)

predict.regression_forest: Predict with a regression forest

Description

Gets estimates of E[Y|X=x] using a trained regression forest.

Usage

# S3 method for regression_forest
predict(object, newdata = NULL,
  linear.correction.variables = NULL, lambda = 0.01,
  ridge.type = "standardized", num.threads = NULL,
  estimate.variance = FALSE, ...)

Arguments

object

The trained forest.

newdata

Points at which predictions should be made. If NULL, makes out-of-bag predictions on the training set instead (i.e., provides predictions at Xi using only trees that did not use the i-th training example).

linear.correction.variables

Optional subset of indexes for variables to be used in local linear prediction. If NULL, standard GRF prediction is used. Otherwise, we run a locally weighted linear regression on the included variables. Please note that this is a beta feature still in development, and may slow down prediction considerably. Defaults to NULL.

lambda

Ridge penalty for local linear predictions

ridge.type

Option to standardize ridge penalty by covariance ("standardized"), or penalize all covariates equally ("identity").

num.threads

Number of threads used in training. If set to NULL, the software automatically selects an appropriate amount.

estimate.variance

Whether variance estimates for hattau(x) are desired (for confidence intervals).

...

Additional arguments (currently ignored).

Value

A vector of predictions.

Examples

# NOT RUN {
# Train a standard regression forest.
n = 50; p = 10
X = matrix(rnorm(n*p), n, p)
Y = X[,1] * rnorm(n)
r.forest = regression_forest(X, Y)

# Predict using the forest.
X.test = matrix(0, 101, p)
X.test[,1] = seq(-2, 2, length.out = 101)
r.pred = predict(r.forest, X.test)

# Predict on out-of-bag training samples.
r.pred = predict(r.forest)

# Predict with confidence intervals; growing more trees is now recommended.
r.forest = regression_forest(X, Y, num.trees = 100)
r.pred = predict(r.forest, X.test, estimate.variance = TRUE)
# }
# NOT RUN {
# }