genER: Generate data using the Erdos and Renyi (ER) model

In yliu433/scZINB: Graphical model estimation for zero-inflated count data

Description

Given the number of vertices, the probability that any two vertices connect, and the sample size, generate the true graph structure and the sample data.

Usage

genER(p, pE, n, effS = NULL)

Arguments

p	number of vertices
pE	probability that two vertices will connect
n	sample size for data generation
effS	the effect size, must be a single number - - default is NULL which generates random coefficients for each covariate from a Uniform(0,1) distribution.

Details

The true graph structure is generated such that the probability of having an edge for any pair of vertices is p_E = d_0/p, where d_0 is average number of edges per node. Then, all graphs with p vertices and d edges have probability of p_E^d(1-p_E)^{\choose(d, 2) -d} to be generated. After we generate the adjacency matrix A, the sample data are simulated from the multivariate normal distribution with mean 0 and covariance

(I - B)^{-1}(I-B^T)^{-1}

where B = the effect size \times the lower triangular matrix of A.

Value

A list with the following components:

A

adjacency matrix of the simulated graph

X

simulated sample data

G

simulated graph as a graphNEL object

Examples

genER(10, 0.1, 20, 0.5)

yliu433/scZINB documentation built on Nov. 30, 2020, 9:07 p.m.