to Petra's homepage
Jožef Stefan International Postgraduate School

2007/2008
Data Mining and Knowledge Discovery
Knowledge Discovery and Knowledge Management in e-Science

Professor: Nada Lavrač
Assistant: Petra Kralj

October 17, 2007 15:00 - 19:00  Lectures (Lavrač)
IPS lecture hall
November 8, 2007
15:00 - 19:00
 Practice (Kralj)
Kolarjeva lecture hall 
November 15, 2007
15:00 - 19:00
 Practice (Kralj)
Kolarjeva lecture hall
November 29, 2007
16:00 - 20:00
 Written exam & seminar proposals
 presentations (Kralj & Lavrač)
Seminarska soba F-5
January 10, 2008
18:00 - 19:00
 Written exam (Kralj)
IPS lecture hall
February 13, 2008
15:00 - 19:00
 Seminar presentations (Lavrač & Kralj)
IPS lecture hall
February 27, 2008
17:00 - 19:00
 Seminar presentations (Lavrač & Kralj)
IPS lecture hall
April 21, 2008
16:00 - 17:30
  Written exam and
  Seminar presentations (Lavrač & Kralj)
IPS lecture hall
 Ask the Jožef Stefan Institute doorkeeper for the location of Kolarjeva lecture hall (Kolarjeva predavalnica).

Materials for the course
    - Lectures, prof. Lavrač
        - lecture notes 17.10.2007 (.pdf)
    - Practice, Petra Kralj
        - practice notes 8.11.2007 (.pdf)
        - entropy and information gain (.pdf)
        - naive Bayes (.pdf)
        - discussion on classification 15.11.2007 (.pdf)
        - practice notes 15.11.2007 (.pdf)
        - evaluating numeric prediction (.pdf)
        - association rules (.pdf)
    - Hands on Weka, Petra Kralj
        - hands on Weka notes - Classification 8.11.2007 (.pdf)
        - hands on Weka notes - Numeric prediction 15.11.2007 (.pdf)
        - hands on Weka notes - Descriptive DM 15.11.2007 (.pdf)
        - datasets (.zip)

Useful links
    - Weka
    - Notepad++

Course requirements:
    - Written exam (1/2 of the final grade)
        - 45 minutes of time
        - 4 tasks (2 computational, 2 theoretical)
        - Literature is not allowed
        - Each student can bring one hand-written A4 sheet of paper and a calculator
    - Seminar (1/2 of the final grade)
        - Oral presentation of seminar proposals (max 4 minutes per student, use slides template,
          file naming convention DM2007-SurnameFirstname.ppt)
        - Register for the seminar presentations as you would usually do for an exam
        - Deliver a written report (printed and electronic copy) in Information Society paper format
          on seminar presentations day (use paper template and guidelines)
        - Oral presentation of seminar results (10 minutes for presentation + 5 minutes discussion,
          use slides template and file naming convention DM2007-SurnameFirstname.ppt

Examples of seminars:
    - Janez Bucik (.pdf)
          Microsoft stock quotes dependency analysis
    - Matej Gašperin (.pdf)
          Case study on the use of data minig techniques in food science using honey samples
    - Valentin Koblar (in Slovene .pdf)
          Napoved menjalnega tečaja ameriškega dolarja na podlagi menjalnih tečajev tujih valut   
       
Ideas for seminars
    - Analyze some data where you are the domain expert, use at least two algorithms
    - Find some interesting data to analyze, possible sources:
        - Statistični urad Republike Slovenije
        - Ljubljanska borza d.d.
        - Banka Slovenije
        - World Health Organization
        - Center for Climatic Research, University of Delaware

Templates
    - presentation template (.pot)
    - paper template (.doc)
    - paper guidelines (.doc)

Link to last year's web page - Data Mining and Knowledge Discovery 06/07
Last update: 20080610