Recognition Labs (AE4B33RPZ)

If you are new to CTU, see checklist for visiting students.

Labs plan

	Date	Topic	Test
1.	25.09.2009	Introduction. SPRT toolbox instalation, work with Matlab, simple example
2.	02.10.2009	Bayesian decision task
3.	09.10.2009	Non-Bayesian tasks - Neyman-Pearson task
4.	16.10.2009	Non-Bayesian tasks - the minimax task	5pts
5.	23.10.2009	Maximum likelihood estimate
6.	30.10.2009	Non-parametrical estimates - Parzen windows	4pts
7.	06.11.2009	Linear classifier - Perceptron
8.	13.11.2009	Support Vector Machines
9.	27.11.2009	Support Vector Machines II	8pts
10.	04.12.2009	AdaBoost
11.	11.12.2009	Clustering - k-means	6pts
12.	18.12.2009	unfinished or bonus tasks

There will be a short test at the beginning of some labs, maximum number of points is given in the table. The questions in the tests will refer to the preceeding lectures.

Requirements to obtain the credit ("zápočet")

• Regular attendance. Only 1 absence is allowed.
• Delivering finished assignment at each lab. Need to be done already during the labs!
• Attending of the tests at the beginning of *-labs. The points from all tests but the worst one will count, so one bad result will not influence your mark. Omitting the worst test applies only if you attended all of the tests!

Compensations

• In the last resort cases (well justified) you will be allowed to compensate for one undelivered assignment or absence at a test.
• The conditions of the compensation will be dealt individually; nevertheless the compensation will in general require more effort than absolving the labs or the test.

Labs evaluation

• The gained points from the labs will contribute to the final classification by 40%.
• The points consist of: test points and bonus points.
• Maximum points for each test are specified in the labs plan table.
• Each finished and delivered bonus task is worth 1-2 points.

Links

• Statistical Pattern Recognition Toolbox
• Matlab for CTU students