BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Constructing adaptive interference-reduced Wigner-Ville spectral e
 stimators of non-stationary time series - Freyermuth\, J-M (Katholieke Uni
 versiteit Leuven)
DTSTART:20140116T113000Z
DTEND:20140116T121500Z
UID:TALK49984@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:Co-authors: Florent Autin (Universit d'Aix-Marseille 1)\, Gerd
 a Claeskens (KULeuven)\, Rainer von Sachs (UCLouvain) \n\nIn this talk we 
 propose estimators of the time-frequency spectrum of a (zero mean) non-sta
 tionary time series with second order structure which varies across time. 
 It is obtained by smoothing the empirical Wigner-Ville (WV) spectrum (Mart
 in and Flandrin\, 1985) which is a highly localized time-frequency spectru
 m. Using the empirical WV avoids prior time-frequency segmentation (such a
 s for the segmented periodogram (Schneider and von Sachs\, 1996)) neverthe
 less it suffers from low and heterogeneous signal-to-noise ratios and from
  severe interferences. In addition\, the associated time-frequency spectru
 m is best modeled as an anisotropic object with locally varying smoothness
  in both time and frequency directions (Neumann and von Sachs\, 1997). All
  this make smoothing very challenging. Our approach is to project the empi
 rical WV data onto a specifically designed hyperbolic wavelet basis (Autin
  et al\, 2013) and to use a tree-structured thresholding (Autin et al\, 20
 11\, 2013) under co nstraints inspired notably by the Heisenberg's uncerta
 inty principle. Such approach is expected to ensure an adaptive time-frequ
 ency representation and to reduce the cross-interferences of the WV spectr
 um.\n\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR