JSC

Jack Storror Carter

I am a EUTOPIA-SIF Marie Sklodowska-Curie COFUND Fellow at Universitat Pompeu Fabra supervised by Prof. David Rossell. I am part of the Statistics group in the Department of Economics and Business and I am also affiliated with the Barcelona School of Economics. My main research interests are in statistical methodology and Bayesian statistics with a focus on graphical models. In my current position I am working on decoupling the concepts of dependence and dispersion in the selection of Gaussian graphical models.

I previously held post-doctoral positions in the Dipartimento di Matematica at the Università di Genova, working with Prof. Eva Riccomagno on the development of methodology for staged tree models, and in the Statistics department at the Univeristy of Warwick. During my PhD, under the supervision of Prof. Jim Smith at the University of Warwick, my research focused on penalised likelihood methods and prior distributions for sparse Gaussian precision matrix estimation and Gaussian graphical model selection. My main contribution in this time was a new penalised likelihood method - the partial correlation graphical LASSO - which provides scale invariant estimation of a Gaussian precision matrix.

Software

A (currently untested) R package for calculation of the partial correlation graphical LASSO is now available via github. The repository is available here and can be installed in R using

devtools::install_github('JackStorrorCarter/PCGLASSO')

Also check out the cegpy python package (which I played a very minor role in updating and maintaining) for selection, estimation and drawing of staged tree models.

Publications

Published papers

Preprints

Thesis