An Accessible Derivation of Linear Regression

The math behind the model, from additive assumptions to pseudoinverse matrices

Technical disclaimer: It is possible to derive a model without normality assumptions. We’ll go down this route because it’s straightforward enough to understand and by assuming normality of the model’s output, we can reason about the uncertainty of our predictions.

This post is intended for people who are already aware of what linear regression is (and maybe have used it once or …

assumptions derivation linear linear regression machine learning math normality optimization reason regression route thoughts-and-theory uncertainty