BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Sanitization for sequential data - Grigorios Loukides (King's Coll
 ege London)
DTSTART:20160929T143000Z
DTEND:20160929T153000Z
UID:TALK67710@talks.cam.ac.uk
CONTACT:INI IT
DESCRIPTION:Organizations disseminate sequential data to support applicati
 ons in domains ranging from marketing to healthcare. Such data are typical
 ly modeled as a collection of sequences\, or a series of time-stamped even
 ts\, and they are mined by data recipients aiming to discover actionable k
 nowledge. However\, the mining of sequential data may expose sensitive pat
 terns that leak confidential knowledge\, or lead to intrusive inferences a
 bout groups of individuals.   &nbsp\;  In this talk\, I will review the pr
 oblem and present two approaches that prevent it\, while retaining the use
 fulness of data in mining tasks.   The first approach is applicable to a c
 ollection of sequences and sanitizes sensitive patterns by permuting their
  events. The selected permutations avoid changes in the set of frequent no
 n-sensitive patterns and in the ordering information of the sequences. The
  second approach is applicable to a series of time-stamped events and sani
 tizes sensitive events by deleting them from carefully selected time point
 s. The deletion of events is guided by a model that captures changes to th
 e probability distribution of events across the sequence.  &nbsp\;
LOCATION:Seminar Room 2\, Newton Institute
END:VEVENT
END:VCALENDAR