BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Efficient entropy-based detection of change-points in streaming da
 ta - James Cruise (Heriot-Watt University)
DTSTART:20110606T133000Z
DTEND:20110606T143000Z
UID:TALK31077@talks.cam.ac.uk
CONTACT:Elena Yudovina
DESCRIPTION:It is well-known that the entropy of an unknown stationary sou
 rce can be consistently estimated using an estimator based on the lengths 
 of long repeated sections of text. I will discuss a method for detecting c
 hange\npoints in data sources based on similar information theory-based qu
 antities. I will present the results of some simulations and theoretical r
 esults based on properties of the typical set which justify how successful
  this method can be.\n\nJoint work with Oliver Johnson\, Dino Sejdinovic\,
  Christophe Andrieu\, Ayalvadi Ganesh and Robert Piechocki
LOCATION:MR12\, Centre for Mathematical Sciences\, Wilberforce Road\, Camb
 ridge
END:VEVENT
END:VCALENDAR