Pattern Classification and Analysis (BBL 514E)

Lecturer

Zehra Cataltepe,  (cataltepe@itu.edu.tr, office: 3301, 0-212-285-3551)

Schedule

Monday 10:00-13:00 (BBL 514E)

Classroom

ENGEL (Egitim Niteligini Gelistirme) Odasi, Computer Engineering Dept. 

Office hour

Monday 13:00-14:00

Announcements

  • Midterm exam: will be in room 2102 between 10am-12noon on monday November 5. It is a books/notes closed exam.
  • Projects announced! Please see the ninova site for more information.
  • Welcome to class! Please email me your email address so that add your name to the ninova pages for the class.

Assignments

  • (due sept 17) Read the first chapter of Introduction to Machine Learning by Alpaydin (EA).
  • (due sept 17) Skim through  Introduction to Probability by -- Charles Grinstead and Laurie Snell.

Course Goals

Gain knowledge and practical experience in pattern recognition/machine learning. Learn practical skills and analytic background for building pattern recognition/machine learning applications.

Prerequisites                                                        

Probability, statistics, linear algebra. Ability/(will to learn) to use MATLAB to do the programming assignments. 

Tentative Course Outline

(Homework and exam dates may change during the term. Please double check!)

Week

Contents

Slides/Links (Old slides will be updated as class proceeds)

Week 1

Sept 17

Introduction, mathematical preliminaries

HW1 handed out

Introduction

Probability and Statistics (by R. Gutierrez-Osuna)

Linear Algebra  (by R. Gutierrez-Osuna)

 

Matlab Exercises (matlabExercises.m, meanStd.m)

 

HW1 (due Oct 15, 2007)

 

Week 2

Sept 24

Supervised Learning

 

Supervised Learning

 

Week 3

Oct 1

Bayesian Decision Theory

 

Bayesian Decision Theory

 

Week 4

Oct 8

Parametric Methods

Parametric Methods

Week 5

Oct 15

Multivariate Methods

HW1 due

 

Multivariate  Methods

 

Week 6

Oct 22

Dimensionality Reduction

Dimensionality Reduction

 

Week 7

Oct 29

NO CLASS: 29 Ekim Cumhuriyet Bayrami

 

Week 8

Nov 5

 

MIDTERM

 

 

 

 

Week 9

Nov 12

Clustering

HW2, 3, 4 handed out

 

Clustering

 

Week 10

Nov 19

Nonparametric Methods

Nonparametric methods

 

Week 11

Nov 26

Decision Trees

 

Decision trees

 

Week 12

Dec 3

Multilayer Perceptrons

 

Multilayer Perceptrons

 

Week 13

Dec 10

Linear Dicrimination

SVMs (Support Vector Machines)

HW2 due

Linear discrimination

 

Week 14

Dec 17

Hidden Markov Models

HW3 due

 

HMMs

sample alpha and beta calculation

Week 15

Dec 24

Assessing and Comparing Classification Algorithms

Combining Multiple Learners

Classifier Assessment and Comparison

Classifier Combination

Week15

Jan 21

9am-1pm

HW4 due

 

PROJECT PRESENTATIONS

Each presentation should take 20 mins, (15 mins presentation + 5 mins for questions and answers). 

 

Grading

Midterm exam

(closed book and notes)

1

%25

Homeworks

4

%10

Project

1

%25

Class participation

(at the instructors discretion)

1

%10

References

  • [EA] Ethem Alpaydn, Introduction to Machine Learning (Adaptive Computation and Machine Learning), The MIT Press, 2004
  • [DHS] Pattern Classification, 2nd Edition, Richard O. Duda, Peter E. Hart, and David G. Stork, 2000, Wiley
  • [TM] Machine Learning, Tom Mitchell, 1997 McGraw Hill.
  • Matlab Introductory Material

From Projects to Conferences:

I encourage students to prepare conference papers out of their projects. In the previous years, the following papers from our Machine Learning/Pattern Recognition classes were accepted to conferences:

  • (2007)  M. Turan and Z. Cataltepe, "Clustering and Dimensionality Reduction to Determine Important Software Quality Metrics", ISCIS 2007, Ankara, Turkey.
  • (2007) Z. Cataltepe, E. Aygun, A. Filiz, O. Keskin, C. Komurlu and Y. Altunbasak, "Dimensionality Reduction for Protein Function Prediction", poster at Automated Function Prediction(AFP)/Biosapiens Joint Special Interest Group Meeting, Vienna, Austria July 2007.
  • (2007)  Z. Cataltepe, H. M. Genc, T. Pearson, "A PCA/ICA Based Feature Selection Method and its Application for Corn Fungi Detection", Eusipco (European Signal Processing Conference) 2007, 3-7 September, Poland.
  • (2007)  H. M. Genc, Z. Cataltepe, T. Pearson, "A New PCA/ICA Based Feature Selection Method/Yeni Bir Temel/Bağmsz Bileşen Analizi(TBA/BBA) Tabanl znitelik Seme Yntemi", SIU 2007, Eskisehir, Turkey.
  • (2007)  E. Karatoprak, S. Seker, Z. Cataltepe and T. Sengler, "Using Bayes Decision Rule for Motor Fault Detection/Bayes Karar Verme Kuramnn Motor Arza Tansnda Kullanlmas", SIU 2007, Eskisehir, Turkey.

Upcoming Conferences in 2007/2008: