BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Structured Dynamic Graphical Models &amp\; Scaling Multivariate Ti
 me Series Methodology - Professor Mike West (Statistical Science\, Duke Un
 iversity)
DTSTART:20160721T100000Z
DTEND:20160721T110000Z
UID:TALK66819@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:I discuss some of our recent R&D with dynamic statistical mode
 ls for multivariate time series forecasting that represents a shift in mod
 elling approaches in response to the coupled challenges of scalability and
  model complexity. Building “simple” and computationally tractable mod
 els of univariate time series is a starting point. Decouple/Recouple is an
  overlaid strategy for coherent Bayesian analysis: That is\, “decouple
 ” a high- dimensional system into the lowest-level components for simple
 /fast analysis\; and then\, “recouple”– on a sound theoretical basis
 – to rebuild the larger multivariate process for full/formal/coherent in
 ferences and predictions.\nI discuss Bayesian dynamic dependency networks 
 (DDNs) and the broader class of simultaneous graphical dynamic linear mode
 ls (SGDLMs) that define a framework to address these goals. Aspects of mod
 el specification\, fitting and computation include importance sampling and
  variational Bayes methods to implement sequential analysis and forecastin
 g. Studies in financial time series forecasting and portfolio decisions hi
 ghlight the utility of the models. The advances in Bayesian dynamic modell
 ing– and in thinking about coherent and implementable strategies for sca
 lability to higher-dimensions (i.e. to “big\, dynamic data”)– are ni
 cely exemplified in these contexts.\n\nAspects of this talk represent rece
 nt joint work with: Zoey Zhao\, 2013 PhD at Duke University\, now at Citad
 el llc\, Chicago\; Lutz Gruber\, 2015 PhD at the Technical University of M
 unich\, now at Quantco\, Cologne\; and Meng Amy Xie\, 2012 BS at Duke Univ
 ersity\, and current PhD student in Statistical Science at Duke.\n\n
LOCATION:James Dyson Building Meeting Room on the Ground Floor
END:VEVENT
END:VCALENDAR