Building machine learning methods from basic principles is a great way to
understand the mathematical and algorithmic intuitions behind the methods. Yes
scikit-learn and such are quicker and easier (and generally faster), but it's
always good to have an understanding of what is going on "under the hood".
This repository will be my examples of a variety of machine learning methods and algorithms from "scratch" (i.e., using minimal or base/common libraries). Check back, as I'll be expanding the list as I have time and as I'm exploring more on my own.
- Simple
$k$ -means clustering by Lloyd's algorithm (Jupyter Notebook, code) - ISOMap dimensionality reduction (coming soon)
- Random Forest (coming soon)