Improved estimation of normalizing constants from MCMC, with application to classical analysis of nonlinear state-space models

Improved estimation of normalizing constants from MCMC, with application to classical analysis of nonlinear state-space models

Neyman Seminar
Feb 24, 2010, 04:00 PM - 05:00 PM | 1011 Evans Hall | Happening As Scheduled
Perry de Valpine, Department of Environmental Science, Policy & Management, UC Berkeley (Speaker)
State-space time-series models for ecological population dynamics (or other applications) combine a latent stochastic model for the dynamics of the true population and a stochastic observation model relating data to the unknown true states. I will begin by introducing applications of nonlinear and/or non-Gaussian state-space models for fisheries, wildlife, and arthropod...