Statistical Science


Computational Statistics

Theme Overview

This theme is concerned with advancing the theory, methodology, algorithmic development and application of simulation based approaches, such as Markov Chain Monte Carlo, to statistical inference.

Theme Members

Ricardo Silva* (Theme Lead)Graphical models; Markov chain Monte Carlo; variational methods
Petros Dellaportas*Markov chain Monte Carlo
Christian HenningClassification; cluster analysis; data visualisation; model selection; multivariate data analysis
Ioanna Manolopoulou*Bayesian statistics; diffusion models; mixture modelling; state-space models
Giampiero Marra*Numerical optimisation; penalised likelihood
Yvo Pokern*High-dimensional Gaussian Markov random fields; inference for partially observed and hypoelliptic diffusion processes; nonparametric inference for stochastic differential equations; sequential Monte Carlo methods
Jinghao Xue*Bootstrapping; data mining; ensemble learning

Other members: Alexandros Beskos*, Serge Guillas*, Franz Király*.

*These theme members are currently accepting applications for PhD supervision

Current and Recent Externally Funded Projects

  • A multicriterion approach for cluster validation, £98k, EPSRC EP/K033972/1, Jun 2013 - May 2017, PI: Hennig.
  • Information geometry for Bayesian hierarchical models, £238k, EPSRC EP/K005723/1, Mar 2013 - Mar 2016, PI: Byrne.
  • Learning Highly Structured Sparse Latent Variable Models, £100k, EPSRC EP/J013293/1, Oct 2012 - Dec 2013, PI: Silva.
  • Sequential Monte Carlo methods for applications in high dimensions, £99k, EPSRC EP/J01365X/1, Jul 2012 - Jun 2013, PI: Beskos.