Cruising The Simplex: Hamiltonian Monte Carlo and the Dirichlet Distribution

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 demonstrate a series of transformations that reshape the canonical Dirichlet distribution into a form much more amenable to MCMC algorithms.

5 pages, 0 figures

References in corpus (1)

• MCMC using Hamiltonian dynamics

Cited by in corpus (14)

• How to Measure Galaxy Star Formation Histories II: Nonparametric Models
• Nested Sampling Methods
• Unsupervised Unmixing of Hyperspectral Images Accounting for Endmember Variability
• Comparing the Galactic Bulge and Galactic Disk Millisecond Pulsars
• The DESI PRObabilistic Value-Added Bright Galaxy Survey (PROVABGS) Mock Challenge
• Information-theoretic metrics for Local Differential Privacy protocols
• New constraints on the up-quark valence distribution in the proton
• The effect of kick velocities on the spatial distribution of millisecond pulsars and implications for the Galactic center excess
• PartonDensity.jl: a novel parton density determination code
• A Bayesian model for sparse graphs with flexible degree distribution and overlapping community structure
• Robust and fast parameter estimation for gravitational waves from binary neutron star merger remnants
• Markov Chain Monte Carlo Methods for Estimating Systemic Risk Allocations
• Improving Frequency Estimation under Local Differential Privacy
• An Efficient Scheme for Sampling in Constrained Domains