BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Using a GLM with spatial random effects to model fractures in anta
 rctic iceshelves  - Sergio Bacallado\, University of Cambridge
DTSTART:20200602T100000Z
DTEND:20200602T113000Z
UID:TALK142135@talks.cam.ac.uk
CONTACT:Jonathan Rosser
DESCRIPTION:Emetc et al. (The Cryosphere\, 12\, 3187–3213\, 2018) recent
 ly gathered a dataset of >100\,000 locations on antarctic iceshelves class
 ified according to whether or not a satellite image shows a fracture. The 
 purpose of their study is to model the statistical relationship between th
 e outputs of a fluid model for the iceshelf and the probability of fractur
 es within it. There is great interest in this type of analysis\, as physic
 s-based simulations of iceshelves with different types of climate forcing 
 could potentially be used to predict the effect of climate change on the r
 isk of iceshelf collapse. Emetc et al. propose a logistic regression model
  with variable selection\, which is applied to different regions within th
 e iceshelves separately\, from which they derive consensus estimates of re
 gression coefficients. \n\nIn this talk\, I will discuss several issues wi
 th the analysis of Emetc et al. and propose an alternative binomial regres
 sion model with spatially correlated random effects. Approximate Bayesian 
 inference is done using a technique proposed by Hensman et al. (NIPS proce
 edings\, 2015) which applies MCMC to sample a sparse gaussian process vari
 ational approximation of the posterior. The technique is scalable to datas
 ets with hundreds of thousands of points. I will show an example with ices
 helf fracture data. \n\nJoint work with Marko Closs and Nick Barrand at th
 e University of Birmingham.\n
LOCATION:https://ukri.zoom.us/j/92946956029
END:VEVENT
END:VCALENDAR