Instructor    (the one...the only...the merciless…)

                                 Dr. Robi Polikar

Office: 136 Rowan Hall

Phone: 256 5372

Office Hours: Open door policy, as always

Class Meeting: Tuesdays & Thursdays —15:15-16:45




Syllabus with Tentative Schedule: Download  syllabus

This Fall in Pattern Recognition

Bayesian classifiers, discriminant analysis, non-parametric density estimation, Parzen windows, K-nearest neighbor classifiers, probabilistic neural network, support vector machines, kernel methods, the multilayer perceptron and radial basis function neural networks, decision trees, unsupervised clustering, ensemble classifiers, boosting, AdaBoost, and Learn++, incremental learning, feature selection, data fusion, confidence estimation, nonstationary learning and more.

Required Texts

1. Pattern Classification 2/e, Duda, Hart & Stork (classic text)

2. Pattern Recognition & Machine Learning—Bishop (comprehensive)

Reference Materials

1. Introduction to Machine Learning—Alpaydin (very introductory)

2. Combining Pattern Classifiers, Kuncheva (Very readable book…!)

3. Neural Networks and Learning Machines 3/e, Haykin (comprehensive)

4. Pattern Recognition 4/e: Theodoridis & Koutroumbras (comprehensive)

5. Learning from Data—Cherkassky, Wiley, 2007.

Homework: Weekly reading and written homeworks will be assigned. Written homeworks will be graded, and you will be quizzed on reading assignments. Homeworks are im-por-tant! Late submissions are not accepted. There will also be a final project. The final project will constitute of an innovative application of pattern recognition on a topic of your interest. A formal project report will be submitted. Graduate students are expected to submit their project work to a (preferably IEEE) conference.

Exams: Two midterms, quizzes, oral reviews.

Grading: HW 25% Midterm 25% Quizzes 10% Project 30% Prof. conduct 10%

Fall  2009



Introduction to & Advanced Topics


Pattern Recognition

ECE 455 / 555