Seagle's winning paper is entitled "Bayesian Melding of the Dead-Reckoned Path and GPS Measur ements for an Accurate and High-Resolution Path of Marine Mammals”.
The ASA’s announcement states in part: “Dear Seagle: Congratulations! You were selected as one of the four award winners for the ENVR paper competition -- out of a huge field of 31 papers!”
The award includes a cash prize of $1500 to cover Seagle’s travel to 2015 JSM conference in Seattle, apart from the substantial recognition in a field of statistics that is growing rapidly in scope and size.
The paper, which represents part of Seagle’s PhD research thesis, is co-authored by Jim Zidek, Brian Battaille, and Andrew Trites (the latter two being from UBC’s Marine Mammal Research Unit). It is a product of an emerging collaboration between that unit and Statistics, one that now involves quite a substantial number of investigators from both groups.
The statistical interest in Seagle’s work lies partly in the immensely long records sea mammals generate in one of their tracks, say, 700,000 autocorrelated entries.
A naive analysis, who might try to invert the 700,000 x 700,000 covariance matrix, would fail! The excellence of Seagle’s paper stems from several ingenious approximations he discovered, making it possible for him to model a track within a Bayesian hierarchical framework....in about 5 minutes! The method promises to have general applicability and to be an important contribution to the new world of Big Data Science.
During his research, Seagle has developed rapport with some of the creatures whose tracks are of the type his methods might handle. If you don’t believe this, check out hishttp://www.stat.ubc.ca/~yang.liu/ website here.
Our heartiest congratulations, Seagle!