Statistical Science


Multivariate and High Dimensional Data

Theme Overview

This theme has a research programme that encompasses both the theoretical and methodological problems encountered when analysing multivariate and high dimensional data. Much of the work in the area is driven by advances in technology in various application fields, where new forms of data with unprecedented levels of heterogeneity and complexity are in a modern setting collected routinely.

Current application problems that the group works on include medical imaging and near-infrared spectroscopy.

Theme members

Tengyao Wang (Theme Lead)CVPR; dimension reduction; image analysis; sparse methods; subspace methods
Petros DellaportasGaussian processes; sparsity
Tom FearnBayesian methods; calibration; classification; multivariate analysis; near infrared spectroscopy
Simon HardenComposite likelihoods
Ricardo SilvaCausal inference; graphical models; latent variable models
Jinghao XueHybrid discriminative-generative classification; imbalanced learning; statistical pattern recognition

Other members: Maria De Iorio, Serge Guillas.

Current and Recent Externally Funded Projects

  • High Dimensional Models for Multivariate Time Series Analysis, £990k, EPSRC EP/I005250/1, Oct 2010 - Sep 2015, PI: Olhede.