Giles Hooker

Giles Hooker
Department of Statistics and Data Science
University of Pennsylvania

Curriculum Vitae


409 Academic Research Building
University of Pennsylvania
265 South 37th Street
Philadelphia, PA 19104-1686

Phone: +1 (215) 898-8222

ghooker@wharton.upenn.edu


I am interested in the interfaces of statistical inference with models both from machine learning, and from applied mathematics. This includes uncertainty quantification in random forests, and ordinary differential equations models. I also work in functional data analysis and in robust statistics. Much of my work is inspired by applications, particularly in ecology, epidemiology, environmental sciences, and medicine.

Research Papers Projects


Of Models and Meanings is a collection of musings on the philosophy of mathematical modeling and what makes a model "interpretable".

Books

Software

fda packages for functional data in R and in Matlab.

CollocInfer is a collection of functions for fitting differential equations to data, I also have an extensive a user manual. See my profiling webpages for somewhat older Matlab code, manuals and demonstrations.

Short Courses

In 2023 I gave a short course on Inference and Interpretation in Machine Learning to the Conference on Applied Statistics in Ireland. A .zip file containing slides, code and data can be found here.

I have also given short courses on functional data analysis. Here is a handout and computer lab from a short course given at IWSM 2010 and JSM 2014 using the fda package.

Songs

For a bit of amusement:

"The Stanford Statistics Songbook: A Musical Tribute". Technical Report, Department of Statistics, Stanford University, 2004.

"Carols for a Cornell Stats-Mass", Technical Report, Department of Statistics and Data Science, Cornell University, 2020.


I am sympathetic to students looking for projects, but do not have the capacity to take on students outside of Penn. I similarly do not have resources to host unadvertised post-docs or visiting scholars, even if self-funded. Due to the volume of requests, I cannot respond to each of these individually. For prospective applications to our undergraduate/PhD programs: I'm happy to discuss the programs if you have questions. However, all admissions are decided centrally and faculty do not sponsor individual applications.