EE 510 DM CLASS SYLLABUS.
DATA MINING AND PATTERN RECOGNITION.
link to Index of Data Mining
Meetings:
OCATE. Mondays and Wednesdays 18:00-19:40.
Teacher:
Marek A. Perkowski, Professor of Electrical Engineering.
Prerequisities:
graduate standing in Electrical Engineering.
MAIN TEXT BOOK:
U. Fayyad, G. Piatetsky-Shapiro, P. Smyth, and R. Uthurusamy,
``Advances in Knowledge Discovery and Data Mining''.
ADDITIONAL TEXTBOOKS AND RESOURCES:
- W. Ziarko, ``Rough Sets, Fuzzy Sets and Knowledge Discovery,''
Springer Verlag, 1994.
- Marek Perkowski, ``Data Mining and Pattern Recognition using Inductive Methods''.
Textbook in preparation.
Will be available from professor, on the chapter-by-chapter basis.
- Zdzislaw Pawlak, ``Rough Sets,''
Kluwer Academic Publishers, 1991.
- Other materials about commercial tools available from professor and his WWW Page.
available from professor.
- Commercial tools information available on WWW Pages; primers, manuals, etc., see packages for EE Department.
available from professor.
Grading Policy:
The grade will be based on projects and class presentations of students.
No exams.
There will be first four short projects/homeworks,
but ulimately every group will work on a larger project.
I will try to make the final project of each group to be related to their previous homeworks.
Each project can be a starting point to a Master Thesis.
Course syllabus:
- What is Data Mining, what is Pattern Recognition?
- Explanation of student projects.
- Fundamentals of data mining.
- Material from Rough Sets by Pawlak - see class handouts Vol. 1.
- Constructive Induction - see class handouts Vol. 2.
- Genetic Algorithms.
- Neural Nets.
- Fuzzy Logic.
- Abduction, Induction and Analogy.
- Advanced concepts in Machine Learning and Knowledge Discovery - see class handouts Vol. 3.
- Hardware Accelerators.
- Applications I: Image Processing - see class handouts Vol. 4.
- Applications II: Data Mining - see class handouts Vol. 5.
- Applications III: Biometric Technologies - see class handouts Vol. 6.
- Applications IV: Banking, Finance, Stock Market - see class handouts Vol. 7.
- Applications V: Knowledge Discovery and Automated Creativity - see class handouts Vol. 8.
Projects to be selected from:
- PROJECT 1: Knowledge Discovery using Constructive Induction.
- PROJECT 2: Data Mining using Generalized Functional Decomposition of Contigency Tables.
- PROJECT 3: Data Mining using Genetic Algorithms with Feedback.
- PROJECT 4: Rough Sets versus Generalized Relation Decomposition:
similarities and differences.
- PROJECT 5: Data Mining using Fuzzy Logic Decomposition and Minimization.
- PROJECT 6: Use of Analogy in Learning.
- PROJECT 7: Inductive Logic Programming. Programming Knowledge Acquisition and Debugging.
- PROJECT 8: Data Mining Machine using Hardware Accelerator based on FPGA technology.
Class will include project related to fast prototyping of digital systems using
DEC-PERLE-1 board, which has 27 Xilinx 3090 chips and 4 Mbytes of static RAM on board.
The entire machine is programmed in a special high level language based on C++.
Individual chips of the machine are programmed using Xilinx tools.
I require at least two people for each project gruoup.
The topics of the possible class projects include:
- PROJECT 9:
Problem of your choice of similiar type (should include one of the following:
complex controller, cellular automata, iterative circuits, pipelined circuits, or systolic circuits)
and of similar complexity.
A project proposed by you will be accepted if it satisfies the following requirements:
- it is not your company proprietary and can be published.
- it is not too easy and/or uninteresting/boring.
- it is not too difficult.
- it solves some problem that was not solved before and is a good illustration
of techniques explained in this class.
CAPSTONE PROJECT 1999
People:
- Tim Brandis - timothy.brandis@orcad.com
- Michael Levy - levym@ee.pdx.edu
- Thang Ta - tat@ee.pdx.edu
- Tu Dinh - tux.dinh.intel.com