School of Computer Science

COMP642 Numerical Estimation

 Course Description

Course Objectives: Estimation theory is a product of need and technology. As a result, it is an integral part of many branches of science and engineering, such as statistics, signal processing, communications, control and navigation. The techniques are used in several areas in computer science, such as machine learning, graphics, image processing, robotics, bioinformatics, and computer vision etc. This course will focus on the design and implementation of efficient and reliable computer algorithms in this area.

Topics:

  • Elements of numerical linear algebra;
  • Basic results of probability theory and statistics;
  • SVD, randomized SVD, and PCA
  • Ordinary least squares estimation
  • Large sparse least squares problems;
  • Generalized least squares estimation;
  • Total least squares estimation;
  • L2-norm regularized least squares estimation;
  • Nonlinear least squares estimation;
  • Maximum-likelihood estimation;
  • Minimum mean square error estimation;
  • Maximum a posteriori estimation;
  • Kalman filtering;
  • L1-norm regularized least squares estimation (LASSO);
  • Integer least squares estimation

    Prerequisites: COMP 350A Numerical Computing or equivalent, MATH 323 Probability Theory or equivalent, a good introductory matrix theory course. COMP 540 Matrix Computations is helpful, but not required.