BEGIN:VCALENDAR
VERSION:2.0
PRODID:researchseminars.org
CALSCALE:GREGORIAN
X-WR-CALNAME:researchseminars.org
BEGIN:VEVENT
SUMMARY:Esther Galbrun (University of Eastern Finland\, Kuopio)
DTSTART:20240424T140000Z
DTEND:20240424T151500Z
DTSTAMP:20260423T021442Z
UID:AAIT/33
DESCRIPTION:Title: <a href="https://researchseminars.org/talk/AAIT/33/">Th
 e minimum description length principle for pattern mining (MDL4PM).</a>\nb
 y Esther Galbrun (University of Eastern Finland\, Kuopio) as part of Semin
 ar on Algorithmic Aspects of Information Theory\n\n\nAbstract\nConsidering
  that patterns express the repeated presence in the data of particular ite
 ms\, attribute values or other discrete properties\, mining patterns is a 
 core task in data analysis. Beyond issues of efficient enumeration\, the s
 election of patterns constitutes a major challenge. The MDL principle\, a 
 model selection method grounded in information theory\, has been applied t
 o pattern mining with the aim to obtain compact high-quality sets of patte
 rns. After introducing some necessary concepts to formalise the pattern mi
 ning task\, we will review MDL-based methods for mining various types of d
 ata and patterns and try to highlight their common characteristics and dif
 ferences. Finally\, we will discuss some of the issues regarding these met
 hods\, and highlight currently active related data analysis problems.\n
LOCATION:https://researchseminars.org/talk/AAIT/33/
END:VEVENT
END:VCALENDAR