A collection of machine learning models implemented from scratch
The models implemented here are for my own learning purposes and that of others.
- Linear Regression from Scratch - published in Towards Data Science
- Neural Networks: why do they work so well? Part I - published in Towards Data Science
A graph of a 7 layer neural netwok.
A 4 layer neural network trained to separate two spirals.
A 5 layer neural network trained to separate three spirals.
A linear regression model fitting a line.
A logistic classifier separates two clusters.
A visualization of all vectors with unit p-norm.
The implementation of these models is inspired by Deep Learning, an outstanding textbook from Bengio and Goodfellow.