rmvnorm.ivent: Simulate from a Gaussian Causal Model

In pcalg: Methods for Graphical Models and Causal Inference

Simulate from a Gaussian Causal Model

Description

Produces one or more samples from the observational or an interventional distribution associated to a Gaussian causal model.

Usage

rmvnorm.ivent(n, object, target = integer(0), target.value = numeric(0))

Arguments

n	Number of samples required.
object	An instance of GaussParDAG
target	Intervention target: vector of intervened nodes. If the vector is empty, samples from the observational distribution are generated. Otherwise, samples from an interventional distribution are simulated.
target.value	Values of the intervened variables. If target.value is a vector of the same length as target, the indicated intervention levels are used for all n samples. If target.value is a matrix of dimension n by length(target), the i-th sample is simulated using the i-th row of the matrix as intervention levels.

Value

If n = 1 a vector of length p is returned, where p denotes the number of nodes of object. Otherwise an n by p matrix is returned with one sample per row.

Author(s)

Alain Hauser ([e-mail])

Examples

set.seed(307)
myDAG <- r.gauss.pardag(5, 0.5)
var(rmvnorm.ivent(n = 1000, myDAG))
myDAG$cov.mat()
var(rmvnorm.ivent(n = 1000, myDAG, target = 1, target.value = 1))
myDAG$cov.mat(target = 1, ivent.var = 0)

pcalg documentation built on May 29, 2024, 5:24 a.m.