You are here: Home Teaching Summer Term 2012 Data Mining Seminar

Data Mining Seminar

Registration Click here to enroll in this course
Course type Seminar
Instructors Stefan Kramer, Prof. Dr., Andreas Karwath, Dr., Michael Geilke, M.Sc., Rui Li, M.Sc., Jörg Wicker, Dipl.-Bioinf.
Time Tuesday, 14:00–16:00
Recurrence weekly from Apr 17, 2012 until Jul 17, 2012
Location 03-424
Credits 4
Directory Link to the directory entry for this lecture
Note:
This course requires registration .

Meeting

 

  • The final meeting will be on Tuesday, July 10. Otherwise there will be no more Seminars.
  • The assignment of the remaining topics to students interested in the seminar will be on Tuesday, April 17 at 14.00 in 03-424.

 

Topics

Date
Speaker
Topic Literature Supervisor
15.5.

Müller

Association Rules

  • Jiawei Han, Micheline Kamber, and Jian Pei, Data Mining: Concepts and Techniques, 3rd edition, Morgan Kaufmann, 2011.

Geilke
15.5.

Kaczmarczyk

Condensed Representation

  • Jean-Francois Boulicaut, Condensed representations for data mining. In: Encyclopedia of Data Warehousing and Mining, J. Wang (Ed.), Idea Group Reference, 2005, pp. 207-211.

Geilke
22.5.

Kepplinger

Graph Mining

  • Managing and Mining Graph Data Charu C. Aggarwal, Haixun Wang Springer-Verlag New York Inc. 2010, Chapter 2

Wicker
29.5.

Babacan

Clustering

  • Jiawei Han, Micheline Kamber, and Jian Pei, Data Mining: Concepts and Techniques, 3rd edition, Morgan Kaufmann, 2011.

Wicker
5.6.

Unger

Support Vector Machines

  • Christopher J.C. Burges, A Tutorial on Support Vector Machines for Pattern
    Recognition, Kluwer, 1998

  • Shigeo Abe, Support Vector Machines for Pattern Classification (Advances in Computer Vision and Pattern Recognition),  Springer, 2005. (Chapter 2 and 3)

Li
5.6.

Tyukin

Kernels

  • John Shawe-Taylor and Nello Cristianini, Kernel Methods for Pattern Analysis, Cambridge, UK, 2004.(Chapter 2 and 3)

  • Bernhard Schlköpf and Alexander J. Smola, Learning with Kernels, MIT Press, 2001.(Chapter 14)

Li

 

Distance Learning

  • Goldberger, J., Roweis, S., Hinton, G., and Salakhutdinov, R. Neighbor-hood Component Analysis. In Proc. of NIPS’04 (2005), pp. 513–520.

  • Weinberger, K., and Tesauro, G. Metric learning for kernel regression. In Proc. of AISTATS’07 (2007).

  • Woznica, A., Kalousis, A., and Hilario, M. Learning to combine distances for complex representations. In Proc. of ICML’07 (2007), pp. 1031–1038. (printed)


12.6.

Weber

Multi-Label Classification

  • G. Tsoumakas, I. Katakis, I. Vlahavas, "Mining Multi-label Data", Data Mining and Knowledge Discovery Handbook, O. Maimon, L. Rokach (Ed.), Springer, 2nd edition, 2010.

Wicker

 

Conditional Random Fields

  • C. Sutton, A. McCallum: An Introduction to Conditional Random Fields for Relational Learning. In: Lise Getoor, Ben Taskar (Hrsg.): Introduction to Statistical Relational Learning. MIT Press, 2006.

  • H. M. Wallach: Conditional random fields: An introduction. University of Pennsylvania, 2004 (Technical Report MS-CIS-04-21).

  • J. Lafferty, A. McCallum, F. Pereira: Conditional random fields: Probabilistic models for segmenting and labeling sequence data. In: Proc. 18th International Conf. on Machine Learning. Morgan Kaufmann, San Francisco, CA 2001, S. 282–289.


19.6.

Bach

Online Learning

  • Avrim Blum: On-line Algorithms in Machine Learning. Online Algorithms 1996: 306-32.

Geilke

 

Tips and Rules

  • Regular participation in the seminar is mandatory for all participants. Those who cannot show up for an important reason , must tell this before the seminar to the supervisor . If you are missing one seminar without an excuse, you will fail in this seminar. The commitment to give a talk is mandatory. A later (possibly short term) reject from the talk will be considered as a false trial.
  • In contrast to "normal" lectures, in the seminar, talks are given  by the students. Thus, the event has other goals than a lecture. T he speaker can learn and practice lecture and presentation techniques in front of a critical audience. Also, he learns t o research independently and to create a (short) scientific work. Therefore, for the success of the seminar the participants themselves are primarily responsible: who gives a poorly prepared lecture, wastes the time of the audience. On the other hand, if the speaker does not get any hints of the success or failure of his lecture from a not concentrated and passive listener.
  • The presentation will last 45 minutes, afterwards, there will be about half an hour for further discussion and comprehension questions . It is expected of each participant that he follows the talk and participates in the discussionat least is able to provide meaningful further questions. It is not expected from the speaker to be able to answer every question . Therefore, it is usually better to admit openly that you do not know the answer to a question than to "fudge" or to speculate.
  • One week after the talk a 5 to 6 page paper of the talk has to be submitted in the ACM style. The paper has to be written in LaTeX.
  • The grade is composed of 1/3 for the talk, 1/3 for the paper and 1/3 for the participation.

 

Checklists:

1. Checklist Preparation

  • Contact your supervisor regarding materials and literature.
  • You are free to search additional materials on your own and use it in your talk if you consulted your supervisor on the materials.
  • If you have any questions during the preparation regarding the topic, do not understand something, or have questions regarding the contents or composition of your talk, please contact your supervisor
  • Start on time!!
  • Exactly two weeks before your talk you automatically have an appointment with your supervisor. On this appointment, the slides have to be done. The supervisor can give you tips and suggestions of change which you can incorporate in your talk.

 

2. Checklist Talk

  • Practice, practice, practice!!
  • Keep an eye on the time (45 minutes!)
  • Speak slowly and clear
  • Look at the audience (not only on the screen!!)
  • You can use the board and/or the overhead projector in addition to your slides

 

3. Checklist Slides

  • The font size has to be at least 18pt
  • Do not use any font with serifs and avoid block quote
  • The space should be used completely but should not be crowded. A good rule of thumb for visual presentations is the 5 plus-minus 2 rule. I.e. 5 plus-minus 2 bullet points per slide.
  • The text of the speaker is presented using bullet points
  • Complete sentences on the slides should only used on important points, e.g. definitions or similar
  • Avoid fancy layouts and flashy colours. Especially a decent frame should be used and give a contrast to the used colours.
  • If at all, use animations moderately.
  • Complex topics can often be explained well using examples.
  • Figures should always have a title and a clear, readable label (e.g. on the axis), so that they can be understood without any explanation of the speaker.
  • Always prefer to recreate a figure by hand than to use a badly scanned image.
  • Sources should be already cited in the presentation using scientific quoting, i.e. title, author, journal, publication date or editor and publisher. Peer reviewed publications should be preferred.
  • The slides should be numbered, so that they can be referred to in questions.

 

4. Checkliste Paper

  • The paper has to written in LaTeX
  • Use the ACM style (http://www.acm.org/publications/submissions/latex_style)
  • In contrast to the slides use complete sentences and literary language.
  • Longer text passages look more uniform in block quote.
  • Check the text for spelling error before submission!
  • Give a number and name to each figure as well as a caption.
  • Do not differ from the given page number (5-6 pages).
  • Write the paper contemporary to the talk and submit it at latest one week after the talk.
  • Write your name on the paper. This makes it easier to map the paper and does not annoy the reviewer.
  • Give all the literature, especially in the citations. The easiest way to do this is bibtex format provided by LaTeX (http://www.bibtex.org/)
News
Wissenschaftsmarkt Posters (Sep 17, 2012)
New Web Page (Feb 16, 2012)
Upcoming Events
Software Engineering - 2nd exam Jun 07, 2013 10:00 AM - 01:00 PM — 05-426
Today's Lectures