Online Program Home
My Program

Abstract Details

Activity Number: 96 - Statistics at NCAR and the Emergence of the Atmospheric Science Statistics Community
Type: Invited
Date/Time: Monday, July 30, 2018 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistics and the Environment
Abstract #326584
Title: Detection of Local Discrepancies Between Two Spatio-Temporal Random Fields
Author(s): Bo Li* and Xianyang Zhang and Sooin Yun
Companies: University of Illinois at Urbana-Champaign and Texas A&M University and Univeristy of Illinois at Urbana-Champaign
Keywords: mixing probability; multiple testing; spatio-temporal data

It is often of interest to compare the characteristics of two spatio-temporal random fields and identify where the difference occurs. For example, comparing the spatially varying trend between climate modeled data and reanalysis data is an important way to evaluate the climate model. We introduce a new adaptive multiple testing procedure to incorporate the spatial dependence to boost power. This is accomplished by modeling the mixing probability as a smooth function over space and using nonparametric approaches within an EM-algorithm to estimate the mixing probability function. We study its theoretical properties including finite sample false discovery rate control, robustness against model misspecification and asymptotic power property.

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

Back to the full JSM 2018 program