BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Problems and algorithms for sequence analysis
DTSTART:20090423T161500
DTSTAMP:20260406T171926Z
UID:af676320b0799450b9866b6c094ab4703329b2c81935d49ac719591f
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr Evimaria Terzi\nSequential data appear in a wide range of d
 iverse applications like telecommunications\, stock-market analysis\, bioi
 nformatics\, text processing\, click-stream mining and many more. In this 
 talk\, I will show how such data motivate new problems and pose novel algo
 rithmic challenges. I will focus on some of these challenges and address t
 hem through the lens of combinatorial optimization.\n\nThe central theme o
 f my talk is sequence segmentation\, i.e.\, the discovery of ``homogeneous
 " segments from input sequences. In the first part of the talk\, I will in
 troduce the notion of segmental groupings and use it to provide comprehens
 ive summaries for large event sequences. Using the Minimum Description Len
 gth principle\, I will define the optimization problem of finding the best
  segmental grouping and then show that it can be solved optimally in polyn
 omial time using a nested dynamic-programming algorithm. For the same prob
 lem\, I will also present faster heuristic algorithms that perform very we
 ll in practice. In the second part of the talk\, I will present a simple a
 nd efficient approximation algorithm for the k-segmentation problem on tim
 e-series data. \nFinally\, I will conclude by showing how the results of d
 ata-analysis algorithms for sequential data can be used as input to new "m
 eta-analysis" algorithms.\n\n\nShort bio: Evimaria Terzi has been a resear
 ch scientist at IBM Almaden Research Center since June 2007. She obtained 
 her PhD from the University of Helsinki (Finland) in January 2007\, under 
 the supervision of prof. Heikki Mannila. Before that\, she got her MSc fro
 m Purdue University (USA) and her BSc from Aristotle University of Thessal
 oniki (Greece). Her research focuses on algorithmic aspects of data mining
  with applications in the analysis of sequential and graph data\, ranking 
 and clustering.\nEvimaria Terzi's home page
LOCATION:BC 410 https://plan.epfl.ch/?room==BC%20410
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR