BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Wavelet-based estimation of the long memory parameter  in Gaussian
  non-gappy and gappy time series - Peter Craigmile\, University of Glasgow
DTSTART:20121019T150000Z
DTEND:20121019T160000Z
UID:TALK39415@talks.cam.ac.uk
CONTACT:Richard Samworth
DESCRIPTION:Knowledge of the long range dependence (LRD) parameter is crit
 ical to\nstudies of fractal and self-similar behavior.  However\, statisti
 cal\nestimation of the LRD parameter becomes difficult when observed data\
 nare masked by short range dependence and other noise\, or are gappy in\nn
 ature (i.e.\, some values are missing in an otherwise regular\nsampling). 
  We investigate estimation of the LRD parameter for\nGaussian time series 
 based upon wavelet variances.  In the non-gappy\ncase\, our least-squares-
 based approach extends and improves upon\nexisting methods by incorporatin
 g correlations between wavelet scales.\nFor the more difficult gappy case\
 , we also develop estimation methods\nby using novel estimators of the wav
 elet variances.  In each case\, we\nprovide asymptotic theory and introduc
 e sandwich estimators to compute\nthe standard errors.  Using Monte Carlo 
 simulations\, we highlight the\nimprovements that are possible over existi
 ng approaches\, and provide\nguidance on practical issues such as how to s
 elect the range of\nwavelet scales.  We consider two applications\; one fo
 r gappy Arctic\nsea-ice draft data\, and another for non-gappy and gappy d
 aily average\ntemperature data collected at 17 locations in south central 
 Sweden.\n\nThis research project is joint with Debashis Mondal\, Ph.D.\, a
 t the\nUniversity of Chicago.\n
LOCATION:MR12\, CMS\, Wilberforce Road\, Cambridge\, CB3 0WB
END:VEVENT
END:VCALENDAR