Transdimensional approximate Bayesian computation for inference on invasive species models with latent variables of unknown dimension

DOI10.1016/J.CSDA.2015.01.002zbMATH Open1468.62041arXiv1310.2888OpenAlexW2106497319WikidataQ56434304 ScholiaQ56434304MaRDI QIDQ1663335FDOQ1663335

Authors: Oksana A. Chkrebtii, Erin K. Cameron, David A. Campbell, Erin M. Bayne 

Publication date: 21 August 2018

Published in: Computational Statistics and Data Analysis (Search for Journal in Brave)

Abstract: Accurate information on patterns of introduction and spread of non-native species is essential for making predictions and management decisions. In many cases, estimating unknown rates of introduction and spread from observed data requires evaluating intractable variable-dimensional integrals. In general, inference on the large class of models containing latent variables of large or variable dimension precludes exact sampling techniques. Approximate Bayesian computation (ABC) methods provide an alternative to exact sampling but rely on inefficient conditional simulation of the latent variables. To accomplish this task efficiently, a new transdimensional Monte Carlo sampler is developed for approximate Bayesian model inference and used to estimate rates of introduction and spread for the non-native earthworm species Dendrobaena octaedra (Savigny) along roads in the boreal forest of northern Alberta. Using low and high estimates of introduction and spread rates, the extent of earthworm invasions in northeastern Alberta was simulated to project the proportion of suitable habitat invaded in the year following data collection.

Full work available at URL: https://arxiv.org/abs/1310.2888

Recommendations

• Bayesian inference for the spatio-temporal invasion of alien species
• Estimation of parameters for macroparasite population evolution using approximate Bayesian computation
• An adaptive sequential Monte Carlo method for approximate Bayesian computation
• Approximate Bayesian computation with differential evolution
• Approximate Bayesian computation and simulation-based inference for complex stochastic epidemic models

zbMATH Keywords

Markov chain Monte Carlolikelihood-free inferencereversible jumpnon-native earthworms

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08) Bayesian inference (62F15) Applications of statistics to biology and medical sciences; meta analysis (62P10)

Cites Work

• Reversible jump Markov chain Monte Carlo computation and Bayesian model determination
• Title not available (Why is that?)
• Constructing Summary Statistics for Approximate Bayesian Computation: Semi-Automatic Approximate Bayesian Computation
• Title not available (Why is that?)
• A note on Metropolis-Hastings kernels for general state spaces
• Title not available (Why is that?)
• Bayesian model choice based on Monte Carlo estimates of posterior model probabilities
• Exact sampling from conditional Boolean models with applications to maximum likelihood inference

Cited In (3)

• A rare event approach to high-dimensional approximate Bayesian computation
• Rapid Bayesian Inference for Expensive Stochastic Models
• Bayesian inference for the spatio-temporal invasion of alien species

This page was built for publication: Transdimensional approximate Bayesian computation for inference on invasive species models with latent variables of unknown dimension

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1663335)