Regularization

Covers matrix based ordinary least squares regression, discretization of Fredholm and Voltarra integral equations, and matrix-based regularization approaches to solve ill-posed problems.

⏻ Learning Objectives

By the end of this module you have learned

  • how to apply matrix based ordinary regression to find multi-linear regression parameters to fit data.

  • how to discretize integral equations to cast them into matrix form.

  • how to analyze matrices to determine the degree to which problems are ill posed.

  • how to apply regularization to invert ill-posed matrix equations.

  • how to apply common approaches to identify the optimal regularization parameter.

Slides and Lecture Notes

Notes: Notes

Notebooks

Notebook: Regression, Inversion, and Regularization

Homework

⌨ Assignments
CC BY-NC 4.0 Markus Petters.