gaussian_pdf function

Proposal density evaluation and simulation

Functions for proposal density evaluation and random generation in MCMC algorithms, in the case where these are Gaussian.

gaussian_pdf(y, x, param)

gaussian_proposal(x, param)

Arguments

• y: Candidate for next move, a vector of dimension $d$
• x: Current position of a chain, a vector of dimension $d$
• param: The proposal parameters, that must contains the $d x d$ variance matrix in param$v.

Details

The Gaussian proposal density $q(y|x)$ used in, e.g., random walk Hastings-Metropolis algorithm RWHM

is the multivariate Gaussian $N(x,v)$ density evaluated at point $y$. Similarly, the Gaussian proposal (next move) is a random draw $y ~ N(x,v)$ when the chain is at position $x$.

Note

These functions are calling multivariate Gaussian density and random generation functions imported from the mixtools package (chosen for efficiency) and wrapped in the format required by the EntropyMCMC package.

Returns

The value of the density, or the random draw, both in dimension $d$

Author(s)

Didier Chauveau.