HACETTEPE UNIVERSITY
Department of Electrical
and Electronics Engineering Course
Syllabus
ELE
673: Pattern Recognition
Credits: 3
Instructors: Assist.
Prof. S. Esen Yuksel, E-mail: eyuksel@ee.hacettepe.edu.tr
Lecture Hours: TBA
Description: This course provides an
introduction to the theory and applications of pattern recognition. Topics
include parametric and
nonparametric classification, feature extraction, clustering. Current medical,
defense and industrial applications.
Textbook: Pattern Classification, R. O. Duda, P. E. Hart
and Stork, Wiley, New York, 2nd Edition, 2001.
Recommended text:
1-
Pattern Recognition and Machine
Learning by C. Bishop, Springer 2006
2-
Data
Mining: Practical Machine Learning Tools and Techniques,
Ian Witten, Eibe Frank and Mark Hall, Morgan Kaufmann Publishers, 3rd
Ed., 2011.
Grading: Homeworks (20%), Project (20%), Midterm (30%), Final (30%)
There will be 4-5
homeworks and one project. For your project, you are required to choose your
own topic at the beginning of the semester.
Attendance: Students are strongly encouraged to attend all classes.
Project:
In this course you will do a substantial project on a
topic that is related to pattern recognition. This project can be: (1) a very
extensive literature search and summary on a particular topic, (2) a good
implementation and evaluation of a known result, in or (3) a small (but
nontrivial) amount of original research. You may work on these projects
individually or in groups of at most two, though if you work in a group, my
expectations will be higher when I grade your project.
You are free to use any programming language and any
opensource toolkit. You can write the codes yourself or use any code that is
available in the public domain. In case you use somebody else's code, you are
required to properly cite its source and know the details of the algorithms
that the code implements.
You are required to submit a project proposal, a final
report written in a conference paper format, and make a presentation during the
mid-term and final weeks. When preparing your report, please use the IEEE
conference format. Tentative schedule of the project is as follows:
- Project proposal (one
week after add-drop): Submit a 1-2 page proposal that describes the
problem you would like to tackle, objective of the study, background of
the problem, related work, etc. Also provide a short list of related
references. (This could contribute to the "introduction" part of
your final report)
- Mid-term progress
presentation and report (presentations will be due one week before the
midterm exam): Make a 5 minute presentation about your progress with the
project, such as the proposed algorithms, hardware/software tools and data
that you plan to utilize, and the evaluation strategies that you plan to
use. Also provide plans for the rest of the semester. Write and submit a mid-term
progress (this could contribute to the "methodology" part of
your final report).
- Final report: Submit
a readable and well-organized report that provides proper motivation for
the task, proper citation and discussion of related literature, proper
explanation of the details of the approach and implementation strategies,
proper performance evaluation, and detailed discussion of the results.
Highlight your contributions and conclusions. Also submit well-documented software with your report.
- Presentation: Make a 10
minute presentation of your work to the class. Each student is expected to
attend all presentations. Each team member should also provide a written
description of her/his own contributions to the project.
I am using this
course to select the students I would like to work with and to give you a
chance to see if you would like to work with me. It is best if you take the class
if you are contemplating working with me for your thesis, or with another
professor on a project related to pattern recognition. These can be excellent
sources to identify your project topic.
COURSE OUTLINE
WEEKS
I. INTRODUCTION ...................................................................................................……........1
I.1 MATHEMATICAL FOUNDATION
II. BAYES DECISION THEORY..............................................................................…….
......1
II.1 Bayes Classier for Continuous Case
II.2 The Gaussian Two-class classifier
II.3 Bayes Classier for Discrete Case
II.4 Error Probability and Receiver Operating Characteristics
III. MAXIMUM-LIKELIHOOD AND BAYESIAN PARAMETER ESTIMATION.......... 1
III.1 Maximum Likelihood Estimation
III.2 Application to Bayesian Classification
III.3 Learning the Mean of Gaussian Density Function
IV. NONPARAMETRIC TECHNIQUES .......................................................................... 2
IV.1 Probability Density Estimation
IV.2 Parzen Windows Estimation
IV.3 k Nearest Neighbor Estimation
IV.4 Nearest Neighbor Rule
IV.5 k Nearest Neighbor Rule
V. LINEAR DISCRIMINANT FUNCTIONS......................................................................
3
V.1 Linear Discriminant Functions and Decision Surfaces
V.2 The Two-Category Case
V.3 Generalized Linear Discriminant Functions
V.4 Relaxation Procedure
V.5 Minimum Square Error Procedure
V.6 Ho-Kashyap Procedure
V.7 Linear Programming Procedure
V.8 Support Vector Machines
VI. UNSUPERVISED LEARNING & CLUSTERING ..................................................... 2
VI.1 Mixture Densities & Identifiability
VI.2 Maximum Likelihood Estimates
VI.3 Applications to Normal Mixtures
VI.4 Unsupervised Bayesian Learning
VI.5 Clustering Techniques & Criterion
VII. ALGORITHM-INDEPENDENT MACHINE LEARNING .....................................
1
VII.1 Bias and Variance
VII.2 Bagging and Boosting
VIII. HYPERSPECTRAL IMAGE PROCESSING
VIII.1 Endmember detection
VIII.2 Abundance estimation
VIII.3 Segmentation and classification
Academic
Integrity: Copying,
communicating, or using disallowed materials during an exam or homework is
cheating. Students caught cheating on a
midterm or final exam will be reported to the campus disciplinary committee. Students
caught cheating on homework will automatically receive a -100 from the homework
and a zero possibly from all homeworks. Any sort of plagiarism will not be
tolerated. This means no copying, no rewording, no paraphrasing, or giving
false or inaccurate information. For more details about plagiarism, please see http://www.plagiarism.org/