## Preliminary list of abstracts

Note: We are using the fantastic MathJax JavaScript library to typeset the mathematics on this web page. You can right-click on a formula to zoom in or to select the rendering engine you like (all under 'Settings'). Under Firefox the default renderer turns out to be MathML (which is quick), but you might prefer the HTML-CSS renderer for more faithful LaTeX rendering. If you encounter any problems with this setup: then please email us!

Exact MCMC inference for Markov switching diffusion models
Keywords: exact algorithms for diffusions; retrospective simulation; MCMC
Wed, 09:25--09:50
• Latuszynski, Krzysztof (Department of Statistics, University of Warwick, United Kingdom)
• Palczewski, Jan (University of Leeds)
• Roberts, Gareth O (University of Warwick)

Consider a stochastic differential equation for $X= X_t$, $t \in [0,T]$, $$dX_t = b(X_t, Y_t, \theta) dt + \sigma(X_t, \theta) \gamma (Y_t, \theta) dB_t.$$ where $B = B_t$ is a Brownian motion, $\theta \in \Theta \subset \mathbb{R}^n$ is a multi-dimensional parameter and the switching process $Y = Y_t,$ $t \in [0,T]$ is a continuous-time Markov process on the state space $\mathcal{Y} = \{1, \dots, m\}$ with the intensity matrix $\Lambda = \{\lambda_{i,j}\}$. Models of this type naturally arise in finance and economics, where $Y$ represents unobservable factors that describe the state of the economy (e.g., a business cycle).

We assume that $X$ is observed at discrete time instances and propose an MCMC algorithm that samples from the exact full posterior $$\pi(X_M, Y, \theta, \Lambda \;| \; X_D) \;\; \propto \;\; \pi_{\theta}(\theta) \pi_{\Lambda}(\Lambda) \pi( Y \, | \, \Lambda) \pi(X_M , X_D \, |\, Y, \theta)$$ including the infinite dimensional sample path of $X$ and avoiding any discretization error. The approach relies on a generalization of the Exact Algorithm of [1,2] and relies on retrospective sampling ideas.

[1] A. Beskos, O. Papaspiliopoulos, G.O. Roberts (2008). A factorisation of diffusion measure and finite sample path constructions. Methodol. Comput. Appl. Probab. 10(1), 85–104.
[2] A. Beskos, O. Papaspiliopoulos, G.O. Roberts, and P. Fearnhead (2006). Exact and computationally efficient likelihood-based estimation for discreetly observed diffusion processes (with discussion). Journal of the Royal Statistical Society B, 68(3), 333-382.