(3-0) Cr. 3. F. Prereq: 424. Mathematical formulation of pattern recognition problems and decision functions, statistical approach, Bayes classifier, pdf estimation, clustering algorithms (supervised and unsupervised), learning algorithms and neural networks, fuzzy recognition systems, feature selection methods, syntactic approach to pattern recognition.
First project on supervised learning with PCA feature reduction.
Project combining different clustering and supervised learning methods and comparison of different pattern recognition systems.
|1||Introduction and Overview, examples||1|
|4||Neural Network Approaches Backpropagation, EM continued; Nonparametric analysis, density estimation, feature reduction||4|
|5||Linear Discriminants, Gradient Descent, Perceptron, LMS, Support Vector MachinesUsing genetic algorithms for clustering; using Nearest Neighbor methods to split a decision space for density estimation||4|
|9||Unsupervised learning and clustering; Methods of cluster analysis, effects of different metrics||5, handout|
|10||Unsupervised learning and clustering; Methods of cluster analysis, effects of different metrics||6|
|11||3/||Unsupervised Clustering Evaluation: Hypothesis testing, internal metrics|
|12||4/3||Tree classification, CART, ID3, C4.5||Papers and website|
|13||4/10||Random Forests, Feature Selection using filters||Breiman Website;|
|14||4/17||Feature Selection using wrappers, Data visualization|
Page last edited 04/10/2006