Participants
François Perron
On some numerical and theoretical aspects of Bayesian inference with a large number of parameters
In this talk we will consider different problems of estimation using
a Bayesian approach and simulation methods for the calculation. We
shall see that some Bayesian approaches can provide simple solutions for
complex problems. However, the calculations are sometimes very complicated and the development of new algorithms is sometimes more
difficult than the solution to the Bayesian problem. In particular,
we shall discuss the idea of ABC ( Approximate Bayes Computation ).
Applications to the problem of estimating copulas will be given.