Markov chain Monte Carlo methods for partially observed diffusions


Séminaire Probabilités & Statistique

5/02/2015 - 14:00 Mr Omiros Papaspiliopoulos (Universitat Pompeu Fabra, Barcelone) Salle 1 - Tour IRMA

I will provide an overview of the state of the art for Bayesian learning of unknown parameters and latent paths using Markov chain Monte Carlo methods in multivariate SDE models. The underlying methodology builds on probabilistic and computational techniques for simulating conditioned stochastic processes.