BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20250921T195222EDT-0981Zwdeb1@132.216.98.100
DTSTAMP:20250921T235222Z
DESCRIPTION:\n	Colloque des sciences mathématiques du Québec\n\nHigh-dimensi
 onal changepoint estimation via sparse projection\n\nRichard Samworth\, Un
 iversity of Cambridge\n\nSite web : http://www.crm.umontreal.ca/Colloques/
 index.html\n\nAbstract:\n\nChangepoints are a very common feature of Big D
 ata that arrive in the form of a data stream. We study high-dimensional ti
 me series in which\, at certain time points\, the mean structure changes i
 n a sparse subset of the coordinates. The challenge is to borrow strength 
 across the coordinates in order to detect smaller changes than could be ob
 served in any individual component series. We propose a two-stage procedur
 e called 'inspect' for estimation of the changepoints: first\, we argue th
 at a good projection direction can be obtained as the leading left singula
 r vector of the matrix that solves a convex optimisation problem derived f
 rom the CUSUM transformation of the time series. We then apply an existing
  univariate changepoint detection algorithm to the projected series. Our t
 heory provides strong guarantees on both the number of estimated changepoi
 nts and the rates of convergence of their locations\, and our numerical st
 udies validate its highly competitive empirical performance for a wide ran
 ge of data generating mechanisms.\n
DTSTART:20161201T203000Z
DTEND:20161201T213000Z
LOCATION:room 1205\, Burnside Hall\, CA\, QC\, Montreal\, H3A 0B9\, 805 rue
  Sherbrooke Ouest
SUMMARY:Richard Samworth\, University of Cambridge
URL:/mathstat/channels/event/richard-samworth-universi
 ty-cambridge-264447
END:VEVENT
END:VCALENDAR