rtellipt function

Simulate elliptical vector subject to a linear constraint

Simulate multivariate Student-t $\boldsymbol{x}$

with location vector mu, scale matrix sigma and df (integer) degrees of freedom subject to the linear constraint $\boldsymbol{\beta}^\top\boldsymbol{x} > 0$. Negative degrees of freedom or values larger than 1000 imply Gaussian vectors are generated instead.

rtellipt(n, beta, mu, sigma, df, delta = 0)

Arguments

• n: number of simulations
• beta: d vector of linear constraints
• mu: location vector
• sigma: scale matrix
• df: degrees of freedom argument
• delta: buffer; default to zero

Returns

an n by d matrix of random vectors