Data Mining Lab Course WS 14/15

 

Type:

Master Lab Course 10 P, IN2106

Ects:

10

Supervisors

Dr. Lothar Richter

Rotation:

weekly meeting of 2 hours, time slot: Wednesday 13-15 (changed)

Rooms:

01.09.014 for the weekly meeting

Language:

English

Announcements:

No class on Jan 7th

The new wiki-logins are ready to be picked up at my office (01.09.041). Already existing student accounts have been extended.

Registration in TUMonline is opened now. Please understand, that is mainly for communication purposes. This does not immediately overrule the matching results. However interested student are welcome since in the first class much less students showed up than there were spots available.

First class on Wednesday Oct 8th at 1 pm
Premeeting takes place on Friday, Jun. 27th, 2014, at 11 o'clock in room 01.09.034

Content

Topics will include (but not be limited to):

  • Selection of interesting data sets
  • Representation of data and information
  • Feature extraction, feature construction, feature selection
  • Data cleaning, discretization
  • Descriptive data mining, finding patterns, clustering
  • Predictive data mining: Finding a prediction task
  • White / black box models
  • Prediction performancs

The emphasis is on getting hands-on experience in the use of data mining and machine learning techniques.

Resources

Course Wiki

Slides

2014-12-17 Preprocessing Performance Measures

2014-11-05 DM Basics

2014-10-08 Orga, Task 1

Hints for next week: Even if you missed the first class you still can participate. Next week every participant has to present briefly a dataset of his/her choice according the instructions given in the slides.