C
Colin Fox
Researcher at University of Otago
Publications - 99
Citations - 2688
Colin Fox is an academic researcher from University of Otago. The author has contributed to research in topics: Markov chain Monte Carlo & Bayesian inference. The author has an hindex of 24, co-authored 96 publications receiving 2355 citations. Previous affiliations of Colin Fox include University of Auckland.
Papers
More filters
Posted Content
Bayesian inference of species trees using diffusion models
TL;DR: In this article, a new and computationally efficient Bayesian methodology for inferring species trees and demographics from unlinked binary markers is described, which uses diffusion models of allele frequency dynamics combined with numerically computing likelihoods of quantitative traits.
Posted Content
Efficiency and computability of MCMC with Langevin, Hamiltonian, and other matrix-splitting proposals
Richard A. Norton,Colin Fox +1 more
TL;DR: This research enables analysis of MCMC methods that draw samples from non-Gaussian target distributions by using AR(1) process proposals in Metropolis-Hastings algorithms, by analysing the matrix splitting of the precision matrix for a local Gaussian approximation of the non- Gaussian target.
Posted Content
Fast sampling in a linear-Gaussian inverse problem
Colin Fox,Richard A. Norton +1 more
TL;DR: This work finds that by efficiently sampling the marginal posterior distribution for hyperparameters, then the full conditional for the deblurred image, it can evaluate the posterior mean faster than regularized inversion, when selection of the regularizing parameter is considered.
A Model For the Propagation of Waves Through the MIZ From a Single Floe Solution
Michael H. Meylan,Colin Fox +1 more
Journal ArticleDOI
Multilevel Delayed Acceptance MCMC
TL;DR: A novel Markov chain Monte Carlo method that exploits a hierarchy of models of increasing complexity to efficiently generate samples from an unnormalized target distribution is developed and shows that the algorithm satisfies detailed balance, hence is ergodic for the target distribution.