Skip to contents

Generate a normal random matrix of covariates/features.

generate_X_gaussian.Rd

Generate an .n x .p normal random matrix with the specified mean and covariance structure.

Usage

generate_X_gaussian(.n, .p, .mean = 0, .sd = 1, .corr = 0, .Sigma = NULL)

Arguments

.n

Number of samples.

.p

Number of features.

.mean

Mean of normal distribution from which to generate data. Can be either a scalar value or vector of length .p. Default is 0.

.sd

Standard deviation of normal distribution from which to generate data. Default is 1.

.corr

Correlation between all pairs of features. Default is 0 for no correlation.

.Sigma

(Optional) .p x .p covariance matrix of the normal distribution from which to generate data. Default is NULL (not used). If provided, the arguments corr and sd are ignored.

Value

A normal random matrix of size n. x .p.

Examples


# Returns 100 x 10 random Gaussian matrix with X_ij ~ N(0, 1)
X <- generate_X_gaussian(.n = 100, .p = 10)

# Returns 100 x 10 random Gaussian matrix with E(X_i) = 0 for all i,
# Var(X_i) = 4 for all i, and Cor(X_i, X_j) = 0.7 for all i != j
X <- generate_X_gaussian(.n = 100, .p = 10, .sd = 2, .corr = 0.7)

# Returns 100 x 2 random Gaussian matrix: X ~ N(0, Sigma), where
# Sigma = [3, .5; .5, 1]
X <- generate_X_gaussian(
  .n = 100, .p = 2, .Sigma = matrix(c(3, .5, .5, 1), nrow = 2, byrow = TRUE)
)