A Flexible Cross-Covariance Function Based on Low-rank Model

Time

-

Locations

Rettaliata Engineering Center, Room 103

Host

Department of Applied Mathematics

Speaker

Juan Hu
Department of Mathematical Sciences, Depaul University
https://csh.depaul.edu/faculty-staff/faculty-a-z/Pages/mathematical-sciences/juan-hu.aspx

Description

Low rank representation for massive spatial data has become popular recently. Model cross covariance structure for multivariate spatial data can be challenging due to the non-negative definite restriction of the covariance function. The commonly used models, such as Linear model of coregionalization, Matern cross covariance function, imposes strong assumptions on marginal spatial process. This project intends to develop a flexible cross-covariance function based on low rank presentation for marginal spatial data. This model can be used in environmental science to study environmental change as well as to predict the weather condition in meteorology.

Event Topic

Computational Mathematics & Statistics

Tags: