Skip to contents

Simulate zero-inflated Poisson log-normal PCA data

Source: R/Simul_fonction.R
Simul.Rd

Generates simulated observations Y under a ZIP-LN-PCA structure from parameters in theta and dimensions in dim.

Usage

Simul(X, theta, dim)

Arguments

X

An \((n \times p) \cdot d\) design matrix stacked so that VectorToMatrix(X %*% B, n, p) produces the mean matrix mu.

theta

List of model parameters with components:

  • B: coefficient matrix (d x 1 or compatible)

  • D: zero-inflation coefficient matrix (d x 1 or compatible)

  • C: loading matrix (p x q)

dim

List of dimensions with entries n, p, d, q.

Value

A list with elements:

  • Y: simulated counts (n x p)

  • W: latent scores (n x q)

  • U: zero-inflation indicators (n x p)

  • Z: Poisson draws before zero-inflation (n x p)

  • Lambda: Poisson intensities (n x p)

  • mu: mean matrix (n x p)

Details

Relies on VectorToMatrix() being available in the package.

Examples

if (FALSE) { # \dontrun{
n <- 5; p <- 4; d <- 3; q <- 2
X <- matrix(rnorm(n*p*d), n*p, d)
theta <- list(B = matrix(rnorm(d)), D = matrix(rnorm(d)), C = matrix(rnorm(p*q), p, q))
dim  <- list(n = n, p = p, d = d, q = q)
out <- Simul(X, theta, dim)
str(out)
} # }