gen_zero_dat: Generates a sample zero-inflated data matrix.

In sqyu/ZiDAG: Directed Graphical Models and Causal Discovery for Zero-Inflated Data

Description

Generates a sample zero-inflated data matrix.

Usage

gen_zero_dat(
  seed,
  gen_para,
  adj_mat,
  n,
  k_mode = 1,
  min_num = 10,
  therun = 1,
  gen_uniform_degree = 1
)

Arguments

seed	A number, the seed passed to set.seed().
gen_para	A string, the parametrization, "abk" or "pms".
adj_mat	A square matrix of 0/1s, the adjacency matrix.
n	A positive integer, the number of samples.
k_mode	One of 0, 1, 2. If k_mode == 1, the k (if gen_para == "abk") or sigmasq (if gen_para == "pms") parameter in each conditional Hurdle distribution is set to 1. If k_mode == 0, the k is set to 1/j (where j is the topological order of the current node) or sigmasq is set to j; if k_mode == 2 they are set in the opposite way.
min_num	A positive integer, the minimum number of 0s/1s in each column. If this is not satisfied, this random generator function will be rerun for up to 1000 times.
therun	A positive integer for internal use. Defaults to 1.
gen_uniform_degree	A positive integer, the maximum degree of the Hurdle polynomials in the conditional Hurdle distributions, defaults to 1. Ignored if parametrization == "abk".

Details

See the simulations section in the paper associated with this package for details.

Value

A generated data, containing

V	A matrix of 0/1s of dimension ncol(adj_mat) x n, equal to Y != 0.
Y	A data matrix of dimension ncol(adj_mat) x n.

Examples

x <- gen_zero_dat(1, "abk", make_dag(3, "chain"), 1000)
x <- gen_zero_dat(1, "pms", make_dag(3, "chain"), 1000)

sqyu/ZiDAG documentation built on Jan. 19, 2021, 4:11 p.m.