Machine Learning 13: Non-Linear Feature Transforms

We discuss shortcomings of linear models for data that is far from linearly separable. We then show how to use non-linear feature transforms to create decision boundaries corresponding to balls, polynomials etc. Finally, we discuss pitfalls when using non-linear transforms based on learning theoretic arguments.

[@sunriseoath]

Watched this video to grok nonlinear transformations of data to model nonlinear processes. This helped a lot of things click for me. Thank you.