Online Program Home
My Program

Abstract Details

Activity Number: 127 - SPEED: Statistical Learning and Data Science Speed Session 1, Part 1
Type: Contributed
Date/Time: Monday, July 29, 2019 : 8:30 AM to 10:20 AM
Sponsor: Section on Statistical Learning and Data Science
Abstract #304695 Presentation
Title: Topological Survival Analysis for the Comparison of Random Fields
Author(s): Hollie Johnson*
Keywords: survival analysis; topological data analysis; statistical topology; climate; spherical correlation; persistent homology

Realisations of simulated climate variables from the CESM Large Ensemble are frequently assumed to be Gaussian iid random fields, with a significant body of work relying on this assumption. Developing concepts from the fields of survival analysis and topological data analysis, we propose a methodology to assess the validity of this assumption with specific application to global wind intensities. We will examine the extent to which the proposed methods are more informative than conventional methods for the comparison of random fields and identification of Gaussianity in data. When applied to global wind realisations, this random field is the 2D sphere embedded in R3, and hence requires modelling of spatial correlation on a sphere. Drawing on work from the field of persistent homology we consider specific topological features, connected components, on a random field. We show that the number of these features differs between fields with different distributions and/or correlation structures and use non-parametric survival models to investigate the rate of emergence of such features. This is achieved through a reformulation of homological ‘births’ and ‘deaths’ as survival events.

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

Back to the full JSM 2019 program