Page 11/09/2017 14:51:32

Project 38: Statistical Modeling of Spatial Extremes

Suitable Majors

Applied Mathematics

Research Area

Statistics​​

Internship Description

The purpose will be to learn about Extreme-Value Theory, and to apply this well-developed theory and related statistical methods to study a real environmental dataset (e.g., extreme rainfall or extreme temperatures or extreme wind speeds, etc., measured at several monitoring stations over a study area). The student will study the `tail' properties (marginal distributions, joint distributions, etc.) of the data, in order to assess the risk of extreme events over the study region.

The methods will be standard methods used in statistics of extremes (GEV, GPD distributions, max-stable processes, etc.). Models will be fitted using a likelihood-based approach.​​

Prerequisites

Strong background in statistics, Experience with the statistical software R, Passion for digging into real environmental data

Deliverables/Expectations

The final product will be in the form of model diagnostics obtained for the real data application (return level maps, probabilities of joint extremes over a spatial region, etc.).​

Other Comments

​Internship dates: 7 July to 14 September​

Division

Computer, Electrical and Mathematical Sciences and Engineering

Faculty Name

Raphael Huser