LAST
UPDATED: 01/08/2012 -
Please see new address
at UCL Department of Statistics.
QRSLab Research Projects
Computational Statistics
Development of novel methodology for sampling, integral estimation and
filtering in stochastic models. Development of algorithms (Markov chain
Monte Carlo MCMC, Adaptive MCMC, Trans-dimensional MCMC, Sequential
Monte Carlo (SMC), non-linear filtering, Likelihood-free sampling
methdology, Annealing and tempering, rare-event simulation).
Bayesian Risk, Commodities, Insurance and Hedge Fund models
Development of novel Bayesian models for Operational Risk, Credit Risk,
Commodities and Non-life insurance claims reserving. This involves
introduction of copula based Bayesian models for correlation
structures; estimation and calibration procedures; joint on-line
parameter estimation and non-linear filtering; likelihood-free
bootstrap procedures; Co-Integrated Vector Autoregression time series
models; multivariate latent factor sde models.