Statistics Seminars: Bayesian parameter estimation for discretely observed diffusions
22 November 2010 15:15 in CM221
In this talk I will describe how MCMC methods can be used to sample from the posterior distribution when estimating parameters of a discretely observed diffusion process. The method described will sample simultaneously from the posterior for the parameters and from the posterior for the (continuous time) diffusion path. The Markov process which form the basis of this MCMC method is a coupled system of a stochastic partial differential equation (SPDE) and a stochastic differential equation (SDE).
Host: Umberto Picchini
Contact email@example.com for more information