BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:High-dimensional latent Gaussian count time series
DTSTART:20231124T151500
DTEND:20231124T170000
DTSTAMP:20260411T053737Z
UID:9fdd3522528eb8257b846aa9d9f196a5f7c59f748136a3201ea9e023
CATEGORIES:Conferences - Seminars
DESCRIPTION:Marie Dueker\, Friedrich-Alexander University\, Germany\nThe f
 ocus of this talk are stationary vector count time series models defined v
 ia deterministic functions of a latent stationary vector Gaussian series. 
 The construction is very general and ensures a pre-specified marginal dist
 ribution for the counts in each dimension\, depending on unknown parameter
 s that can be marginally estimated.  The vector Gaussian series injects f
 lexibility in the model's temporal and cross-sectional dependencies\, perh
 aps through a parametric model akin to a vector autoregression. This talk 
 discusses how the latent Gaussian model can be estimated by relating the c
 ovariances of the observed counts and the latent Gaussian series.  In a p
 ossibly high-dimensional setting\, concentration bounds are established fo
 r the differences between the estimated and true latent Gaussian autocovar
 iances\, in terms of those for the observed count series and the estimated
  marginal parameters. An application of the result is given to the case wh
 en the latent Gaussian series follows a VAR model\, and its parameters are
  estimated sparsely through a LASSO-type procedure.\n\nThe talk is based o
 n joint work with Robert Lund (University of California - Santa Cruz) and 
 Vladas Pipiras (University of North Carolina - Chapel Hill)\n 
LOCATION:MA A3 30 https://plan.epfl.ch/?room==MA%20A3%2030
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR