Online Program Home
My Program

Abstract Details

Activity Number: 619 - Spatial and Spatial-Temporal Statistics
Type: Topic Contributed
Date/Time: Thursday, August 2, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics and the Environment
Abstract #329454 Presentation
Title: Non-Gaussian Translation Processes in Dynamic Space-Time Modeling
Author(s): Robert Richardson*
Companies: Brigham Young University
Keywords: Stochastic Process; Extended Kalman Filter; Extreme Value Modeling

Non-Gaussian translation processes is a method used primarily in engineering to model non-Gaussian stochastic processes. With a strong connection to copulas, this methodology separates the correlation structure of a stochastic process from the marginal distributions of the specific data points, allowing for flexibility in user selection of distributional characteristics. When applied in a multivariate setting to space-time models, the result is a non-linear spatio-temporal dynamic model, that can be fit using an extended Kalman filter. It can be combined with a number of existing parameterizations of linear dynamic model structures allowing these existing methods to be easily extended to include non-Gaussian marginals. The methodology is described along with theoretical properties of the resulting process. It is then applied in an extreme value setting using stable laws as marginal distributions.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program