These lecture notes develop basic least squares solutions to linear systems of equations. Several examples from signal processing are given to illustrate the use of least squares in a variety of problems.

Download the lecture notes: least_squares_SP (pdf file)

This tutorial is also available on the Connexions module.

Download the Matlab demos: LeastSquares_SPdemos.zip (zip file)

- Polynomial approximation
- Linear prediction
- Smoothing
- Deconvolution
- System identification
- Estimating missing data
- Speech de-clipping

Estimating missing samples by least squares (regularization of the energy of the second-order derivative) subject to a data consistency constraint.

Ivan Selesnick

Polytechnic Institute of New York University

Electrical and Computer Engineering

Brooklyn, New York, United States