Course Description:

Machine learning and pattern classification are fundamental blocks in the design of an intelligent system. This course will introduce machine learning and pattern classification concepts, theories, and algorithms. The topics covered in this course include: Bayesian decision theory, linear discriminant functions, support vector machines (SVM), multilayer neural networks, classifier evaluation methods, unsupervised learning and clustering, component analysis (PCA, ICA), genetic algorithm, and classification and regression trees (CART).   

Prerequisites:  All required third year courses.

Format: Lectures, Tutorials, Labs, Assignments (or) Project.

Announcement: This course information and materials are managed through the Ryerson Blackboard Learning Management System. Students are advised to check the announcements regularly in the Blackboard system.

Instructor:

Dr. K. Umapathy
Office: ENG 460
Tel: (416) 979 5000 ext. 4152
Fax: (416) 979 5280
Email: karthi@ee.ryerson.ca
Office hours: Tue & Thu  10am to 11am
(or) by appointment

Teaching Assistant:

TBA

Detailed course description: Course_overview_ELE888_EE8209.pdf

Laboratory Components (Approximate Hours):

1. Bayesian decision theory – (3 hrs)

2. Linear discriminant functions – (3 hrs)

3. Neural networks and algorithm independent machine learning – (3 hrs)

4. Unsupervised learning and clustering – (3 hrs)

Lab hours are flexible; however the Instructor/TA will be present only during the posted lab hours.

Assignment (or) Project Components: 

Students use MATLAB to design and implement their lab projects. Students are also encouraged to use C, C++ or JAVA as programming language tools. Assignment/Project details will be discussed in the class.

Prescribed Text(s):

1. R. O. Duda, P. E. Hart and D. G. Stork, Pattern Classification, 2nd edition, John Wiley & Sons, 2002. ISBN: 0-471-05669-3.

2. Class notes will be provided where applicable.

References:

1. Statistical Pattern Recognition, 2nd edition, A. Webb, John Wiley & Sons.

2. Statistical Pattern Recognition: A Review, Anil K. Jain, Robert P.W. Duin, Jianchang Mao, IEEE Transactions on Pattern Analysis and Machine Intelligence.

3. Other relevant research papers in the literature.

Marking Scheme:

1. Mid-term Exam    - 30%

2. Lab Projects         - 40%

3. Final Exam           - 30%

Mid-term and final exams will be closed-book exams. If necessary the instructor may provide a formula sheet.

Note: All the required reports will be assessed not only on their technical or academic merit, but also on the communication skills of the author as exhibited through the reports.