Theoretical Statistics and Mathematics Unit, ISI Delhi

On Adaptive Rejection Sampling

by Rajeeva L. Karandikar

Adaptive Rejection Algorithm (ARS) is an algorithm to generate samples from a log-concave density. The algorithm, due to Gilks [1], is very efficient. As in MCMC, this requires that the user specify the density upto a normalising constant. However, unlike in MCMC, this algorithm produces i.i.d. samples. The algorithm requires the user to also give a set of points $x_0

isid/ms/2005/09 [fulltext]

Click here to return to Preprints Page