CSCE 666: Pattern Analysis
Spring 2020

Instructor: Ricardo Gutierrez-Osuna
Office: 506A HRBB
Phone: 979.845.2942


Course description: Introduction to methods for the analysis, classification and clustering of high-dimensional data in Computer Science applications. Course contents include density and parameter estimation, linear feature extraction, feature subset selection, clustering, Bayesian and geometric classifiers, non-linear dimensionality reduction methods from statistical learning theory and spectral graph theory, Hidden Markov models, and ensemble learning.

Prerequisites: CPSC 206, MATH 222, MATH 411 (or equivalent) and graduate standing in CPSC, CECN, ELEN, CEEN (or permission of the instructor). Basic knowledge of Linear Algebra, Probability and Statistics: algebra of matrices, geometry of Euclidean space, vector spaces and subspaces, basis, linear independence, linear transformations, eigenvalues and eigenvectors, mean, variance, probability and distributions. Programming experience in a high-level language is required.

Lecture notes

This material is NOT intended to be comprehensive, but rather a SUMMARY of the key concepts covered in the lectures. Consult the syllabus for additional reading material from the textbook, which may be included in the tests.

Lecture slides
Introduction to Pattern Recognition
Review of Statistics and Probability
Linear Algebra and MATLAB
Fourier Analysis
Bayesian Decision Theory
Quadratic Classifiers
Parameter Estimation
Kernel Density Estimation
Nearest Neighbors
Linear Discriminant Functions
Principal Components (GIF)
Fisher's Linear Discriminants
Feature Subset Selection
Advanced Dimensionality Reduction
Mixture Models and EM (MPEG)
Statistical Clustering (MPEG)
Independent Components Analysis
Support Vector Machines
SVMs and Kernel Methods
Kernel PCA/LDA
Discrete HMMs, Viterbi
Baum-Welch and Entropic Training
Ensemble Learning

Homework assignments

Homework #