generate_dataset: Simulates an intervention dataset.

In brielin/inspre: INverse SParse REgression

Simulates an intervention dataset.

Description

Simulates a graph-intervention dataset with corresponding graph.

Usage

generate_dataset(
  D,
  N_cont,
  N_int,
  int_beta = -2,
  graph = "scalefree",
  v = 0.2,
  p = 0.4,
  DAG = FALSE,
  C = floor(0.1 * D),
  noise = "gaussian",
  model = "inspre"
)

Arguments

D	Integer. Number of nodes to simulate.
N_cont	Integer. Number of control samples to simulate.
N_int	Integer. Mean of number of intervention samples to simulate per node.
graph	String. One of 'scalefree' or 'random'. Type of network to simulate.
DAG	Bool. TRUE to ensure the returned graph is a DAG.
C	Integer. Number of fully connected confounding nodes to simulate.
noise	String. Noise model to simulate, currently just "gaussian".
model	Data generating model. One of "inspre" or "dotears".
size	Float. Size parameter for NB distribution for int samples.
int_r2	Float. Mean variance in node explained by intervention.
int_dir	string. 'positive', 'negative', or 'both'. Direction of effect of intervention on node.
v_mode	Float. Mode of the pert distribution.
v_min	Float. Min of the pert distribution.
v_max	Float. Max of the pert distribuion.

brielin/inspre documentation built on Dec. 3, 2023, 4:55 a.m.