BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Wild Binary Segmentation for multiple change-point detection - Fry
 zlewicz\, P (London School of Economics)
DTSTART:20140114T133000Z
DTEND:20140114T141000Z
UID:TALK49864@talks.cam.ac.uk
CONTACT:Mustapha Amrani
DESCRIPTION:We propose a new technique\, called Wild Binary Segmentation (
 WBS)\, for consistent estimation of the number and locations of multiple c
 hange-points in data. We assume that the number of change-points can incre
 ase to infinity with the sample size. Due to a certain random localisation
  mechanism\, WBS works even for very short spacings between the change-poi
 nts\, unlike standard Binary Segmentation. On the other hand\, despite its
  use of localisation\, WBS does not require the choice of a window or span
  parameter\, and does not lead to significant increase in computational co
 mplexity. WBS is also easy to code. We describe two types of stopping crit
 eria for WBS: one based on thresholding and another based on what we call 
 the "extended Schwarz Information Criterion". We provide default recommend
 ed values of the parameters of the procedure and\, in an extensive simulat
 ion study\, show that it offers very good practical performance in compari
 son with the state of the art. We provi de a new proof of consistency of B
 inary Segmentation with improved rates of convergence\, as well as a corre
 sponding result for WBS.\n
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR