EE547 will be oriented to hands-on pattern recognition projects, in the form of programming assignments and larger projects using real-world data from a variety of sources such as bioinformatics and computer security. Course prerequisite is basic calculus and introductory statistics.
Description: (3-0) Cr. 3. F. Mathematical formulation of pattern recognition problems and decision functions, statistical approach, Bayes classifier, probability density function estimation, clustering algorithms (supervised and unsupervised), learning algorithms and neural networks, fuzzy recognition systems, feature selection methods.
|Homework Assignments and paper/algorithm presentations in class 30%|
|Project 1 (Report Only) 25%|
|Final Project (Report and Presentation) 35%|
|Class Participation 10%|
|1||Introduction and Overview, examples||1|
|5||Nonlinear Classifiers; Kernel-based approaches||4|
|6||Feature Selection and Processing, statistical methods, project 1 Face Recognition||5.1,5.2,6.1, handouts,|
|7||Methods of cluster analysis, effects of different metrics||10|
|9||Clustering Function Optimization||12|
|10||Cluster validity and assessment||13|
|11||Feature Selection and Generation||5|
|12||Tree classification, CART, ID3, C4.5||Papers and website|
|13||Random Forests, Feature Selection using filters||Breiman Website;|
Feature Selection using wrappers, Data visualization
The demonstrations presented in this course will be mostly performed in Matlab. However, students are welcome to use any programming or analysis package that they are most comfortable with. All students will be given access to the College of Engineering computer network for their course work and tutorial sessions on how to use Matlab will be held for interested students.
Page last edited 04/10/2006