simple_tree_data: Generate simple tree data.

In banditsCI: Bandit-Based Experiments and Policy Evaluation

Generate simple tree data.

Description

Generates covariates and potential outcomes of a synthetic dataset for a simple tree model.

Usage

simple_tree_data(
  A,
  K = 5,
  p = 10,
  noise_std = 1,
  split = 1.676,
  signal_strength = 1,
  noise_form = "normal"
)

Arguments

A	Integer. Number of observations in the dataset. Must be a positive integer.
K	Integer. Number of arms. Must be a positive integer.
p	Integer. Number of covariates. Must be a positive integer.
noise_std	Numeric. Standard deviation of the noise added to the potential outcomes. Must be a non-negative number.
split	Numeric. Split point for creating treatment groups based on the covariates.
signal_strength	Numeric. Strength of the signal in the potential outcomes.
noise_form	Character. Distribution of the noise added to the potential outcomes. Can be either "normal" or "uniform".

Value

A list containing the generated data (xs, ys, muxs) and the true potential outcome means (mus).

Examples

set.seed(123)
A <- 1000
K <- 4     # Number of treatment arms
p <- 10    # Number of covariates
synthetic_data <- simple_tree_data(A = A,
                                   K = K,
                                   p = p,
                                   noise_std = 1.0,
                                   split = 1.676,
                                   signal_strength = 1.0,
                                   noise_form = 'normal')

banditsCI documentation built on April 12, 2025, 1:42 a.m.