ACADEMICS
Course Details

ELE673 - Pattern Recognition

2024-2025 Fall term information
The course is not open this term
ELE673 - Pattern Recognition
Program Theoretıcal hours Practical hours Local credit ECTS credit
MS 3 0 3 8
Obligation : Elective
Prerequisite courses : -
Concurrent courses : -
Delivery modes : Face-to-Face
Learning and teaching strategies : Lecture, Question and Answer
Course objective : In order to equip the students with the capability to solve real-life problems in pattern recognition, this course aims to teach the following topics to the students: ? basic concepts in pattern recognition, ? basics of statistical decision theory, ? parametric and nonparametric approaches and their differences, ? other techniques used in moders pattern recognition systems, while mainly staying in the context of statistical techniques.
Learning outcomes : Know the basic concepts and approaches in pattern recognition, Know the comparative advantages and disadvantages of different approaches, Apply the techniques and algorithms s/he learnt in the class in real-life applications, Propose realistic solutions to previously unencountered pattern recognition problems, Have the adequate knowledge to follow and understand advanced up-to-date pattern recognition algorithms.
Course content : Basics of pattern recognition: Pattern classes, features, feature extraction, classification. Statistical decision theory, Bayes classifier, Minimax and Neyman-Pearson rules, error bounds. Supervised learning: Probability density function estimation, maximum likelihood and Bayes estimation. Nonparametric pattern reconition techniques: Parzen windows, nearest neighbor and k-nearest neigbor algorithms. Discriminant analysis, least squares and relaxation algorithms. Unsupervised learning and clustering. Other approaches to pattern recognition.
References : Duda R. O., Hart P. E., and Stork D. G., Pattern Classification, 2nd ed., John Wiley and Sons, 2001.; Webb A., Statistical pattern recognition, Oxford University Press Inc., 1999.; Theodoridis S., Koutroumbas K., Pattern recognition, Academic Press, 1999.
Course Outline Weekly
Weeks Topics
1 Basic concepts in pattern recognition
2 Bayesian decision theory, Error integrals, Minimax and Neyman-Pearson rules
3 Discriminant functions for the multivariate normal density, Error bounds for normal densities: Chernoff and Bhattacharyya bounds
4 Bayes decision theory for disrete features, Missing and noisy features
5 Parameter estimation: Maximum likelihood and Bayes estimation, The notion of sufficient statistic
6 Problems of dimensionality, Principle component analysis and Fisher linear discriminant analysis
7 Nonparametric techniques: Parzen windows
8 Nonparametric techniques: nearest neighbor and k-nearest neighbor algorithms, Common metrics used in pattern recognition
9 Midterm Exam
10 Linear discriminant functions and decision regions
11 Gradient descent methods: Perceptron algorithm, relaxation algorithms
12 Least squares algorithm, Support Vector machines
13 Unsupervised learning: Clustering algorithms, k-means clustering, Performance measures in clustering: Minimum variance and scattering criteria
14 General overview of non-statistical pattern recognition techniques, Decision trees, strings and grammar based methods
15 Preparation week for final exams
16 Final exam
Assessment Methods
Course activities Number Percentage
Attendance 0 0
Laboratory 0 0
Application 0 0
Field activities 0 0
Specific practical training 0 0
Assignments 7 35
Presentation 0 0
Project 0 0
Seminar 0 0
Quiz 0 0
Midterms 1 25
Final exam 1 40
Total 100
Percentage of semester activities contributing grade success 60
Percentage of final exam contributing grade success 40
Total 100
Workload and ECTS Calculation
Course activities Number Duration (hours) Total workload
Course Duration 14 3 42
Laboratory 0 0 0
Application 0 0 0
Specific practical training 0 0 0
Field activities 0 0 0
Study Hours Out of Class (Preliminary work, reinforcement, etc.) 14 8 112
Presentation / Seminar Preparation 0 0 0
Project 0 0 0
Homework assignment 7 8 56
Quiz 0 0 0
Midterms (Study duration) 1 10 10
Final Exam (Study duration) 1 20 20
Total workload 37 49 240
Matrix Of The Course Learning Outcomes Versus Program Outcomes
Key learning outcomes Contribution level
1 2 3 4 5
1. Has general and detailed knowledge in certain areas of Electrical and Electronics Engineering in addition to the required fundamental knowledge.
2. Solves complex engineering problems which require high level of analysis and synthesis skills using theoretical and experimental knowledge in mathematics, sciences and Electrical and Electronics Engineering.
3. Follows and interprets scientific literature and uses them efficiently for the solution of engineering problems.
4. Designs and runs research projects, analyzes and interprets the results.
5. Designs, plans, and manages high level research projects; leads multidiciplinary projects.
6. Produces novel solutions for problems.
7. Can analyze and interpret complex or missing data and use this skill in multidiciplinary projects.
8. Follows technological developments, improves him/herself , easily adapts to new conditions.
9. Is aware of ethical, social and environmental impacts of his/her work.
10. Can present his/her ideas and works in written and oral form effectively; uses English effectively.
1: Lowest, 2: Low, 3: Average, 4: High, 5: Highest