Computational methods for fitting statistical models to spatial data

Murali Haran, PhD

Pennsylvania State University,
Pennsylvania, USA

Abstract

Spatial data arise in many subjects, including disease modeling, environmental science, mining and engineering problems. Fitting statistical models to such data can be computationally challenging. I will describe algorithms that efficiently fit statistical models called Gaussian random field models to large spatial data sets. This is based on joint work with Yawen Guan (Statistical and Applied Mathematical Sciences Institute, North Carolina, USA), and John Hughes (University of Colorado-Denver, USA).

Short bio

Murali Haran https://www.niss.org/people/murali-haran is a professor of statistics at Penn State. He obtained his PhD in statistics from the University of Minnesota and his BS in computer science from Carnegie Mellon University. His primary methodological research interests are statistical computing (Markov chain Monte Carlo methods), models for spatial data, and statistical methods for analyzing complex computer models. Much of his work is motivated by applications in the environmental sciences, including climate science research and infectious disease modeling.