This repository contains many ml algorithms coded from scratch. You can find a short description of each project below.
This is an implementation of ridge regression (l2 regularised regression) applied to the Boston housing dataset.
Vanilla logistic regression applied to the banknote authentification dataset. An accuracy of 98.7% percent was achieved.
Simple K nearest neighbors algorithm applied to the mnist dataset. An accuracy of 95.5% percent was achieved.
Naive Bayes classifier applied to imdb, yelp and amazon reviews. An accuracy of 78.3% percent was achieved.