PTOforest: Fit a pollinated transformed outcome (PTO) forest model

In saberpowers/causalLearning: Methods for heterogeneous treatment effect estimation

Description

Fit a pollinated transformed outcome (PTO) forest model

Usage

PTOforest(x, tx, y, pscore = rep(0.5, nrow(x)), num.trees = 500,
  mtry = ncol(x), min.node.size = max(25, nrow(x)/40), postprocess = TRUE,
  verbose = FALSE)

Arguments

x	matrix of covariates
tx	vector of treatment indicators (0 or 1)
y	vector of response values
pscore	vector of propensity scores
num.trees	number of trees for transformed outcome forest
mtry	number of variables to possibly split at in each node
min.node.size	minimum node size for transformed outcome forest
postprocess	logical: should optional post-processing random forest be fit at end?
verbose	logical: should progress be printed to console?

Value

an object of class PTOforest with attributes:

• x: matrix of covariates supplied by function call

• pscore: vector of propensity score supplied by function call

• postprocess: logical supplied by function call

• TOfit: fitted random forest on transformed outcomes

• PTOfit1: TOfit pollinated with treatment-arm outcomes

• PTOfit0: TOfit pollinated with control-arm outcomes

• postfit: post-processing random forest summarizing results

Examples

# Randomized experiment example

n = 100 # number of training-set patients to simulate
p = 10  # number of features for each training-set patient

# Simulate data
x = matrix(rnorm(n * p), nrow = n, ncol = p) # simulate covariate matrix
tx_effect = x[, 1] + (x[, 2] > 0) # simple heterogeneous treatment effect
tx = rbinom(n, size = 1, p = 0.5) # random treatment assignment
y = rowMeans(x) + tx * tx_effect + rnorm(n, sd = 0.001) # simulate response

# Estimate PTO forest model
fit_pto = PTOforest(x, tx, y)
pred_pto = predict(fit_pto, newx = x)

# Visualize results
plot(tx_effect, pred_pto, main = 'PTO forest',
 xlab = 'True treatment effect', ylab = 'Estimated treatment effect')
abline(0, 1, lty = 2)

saberpowers/causalLearning documentation built on May 30, 2019, 8:26 a.m.