Summer School
Advanced Computational Methods for Statistical Inference
Organizers
Eric Moulines, C.P. Robert
Speakers: Peter Clifford, Hans Kuensch, Gareth Roberts, Neil Shephard.
More informations
http://www.tsi.enst.fr/~asseline/cirm/presentation.html
Preregistration
http://www.tsi.enst.fr/~asseline/cirm/preregistration.html
----------------------------------------------------------------------------
----
The summer school is funded by the European Union's TMR network "Statistical
and Computational Methods for the Analysis of Spatial Data.
ERB-FMRX-CT96-0096"
----------------------------------------------------------------------------
----
The purpose of this 5 days summer school (September 17-21, 2001), is to
bring together senior researchers on the topic of recent advances in
computational methods, with applications to complex statistical problems,
like those encountered in econometrics, signal processing and finance, and
to introduce Ph.D. students and junior researchers to these methods. The
school will be organised in order to meet these two requirements, with
courses held in the mornings, advanced talks in the afternoon and
round-tables in the evenings. In fact, new methods are currently appearing
in the field of computational Statistics which go beyond standard MCMC
algorithms, taking advantage of earlier works on importance sampling
(auxiliary variables, stratified sampling) and particle systems, the later
being mainly studied in probability theory and statistical physics so far.
Different simulation methods have also been devised for the analysis of
continuous-time stochastic processes, in particular in connection with
stochastic differential equations. The applications of these methods in
various fields are also undergoing a rapid growth: see for instance the use
of particle filters in target tracking and image analysis, advanced signal
processing, quantitative finance and econometrics.
Preliminary Program
The format of the school consists in 3 courses of 5 hours each, plus
afternoon tutorials, on
G. Roberts (Univ. of Lancaster): Advanced methods in Markov Chain
Monte-Carlo; (tutor: S. Jarner)
P. Clifford (Jesus College, Oxford) and H. Kuensch (ETH, Zurich): Particle
filter methods (tutor: Ch. Andrieu)
N. Shephard (Nuffield College, Oxford): Computational methods for hidden
markov processes and switching autoregressive processes (tutor: M. Pitt)
Contributed sessions will be organised on
Sequential methods for estimation, filtering
Applications
Latent variables models
Stochastic volatility models and Monte-Carlo methods in Finance
Particle filters and other importance sampling methods
Current trends in MCMC: jump sampler, birth-death processes, perfect
sampling, etc.
The slides of the courses and handouts will be made available at the time of
the conference. Round tables and poster sessions will be organized to
increase the exchanges between the participants to the school.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|