Year: 2012
Type: conference-paper
Abstract: Due to its constrained support, the Dirichlet distribution is uniquely suited to many applications. The constraints that make it powerful, however, can also hinder practical implementations, particularly those utilizing Markov Chain Monte Carlo (MCMC) techniques such as Hamiltonian Monte Carlo. I introduce a series of transformations that reshape the canonical Dirichlet distribution into a form more amenable to efficient MCMC algorithms and demonstrate its utility with a few examples.
Source: AIP conference proceedings
DOI: 10.1063/1.3703631
Author: Michael Betancourt
Institution: Massachusetts Institute of Technology
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2013892431