This repository is a replication of my work for an Introductory Machine Learning coursework at The University of Edinburgh, which I originally completed in MATLAB. I wanted to learn Python and get familiar with Numpy, so here we are!
The goal of this assignment was to classify, as accurately as possible, handwritten characters taken from the EMNIST dataset (https://www.nist.gov/itl/iad/image-group/emnist-dataset). Each character image was represented as 28-by-28 pixels in gray scale, being stored as a row vector of 784 elements (28 ร 28 = 784). This coursework was completed using MATLAB 2015.
Using K-NN classification.
Using Naive Bayes classification with multivariate Bernoulli distributions.
Using Bayes classification with Gaussian distributions, where each class (character) is modelled with a multivariate Gaussian distribution.
Same as done in 3A, with the addition of a K-means clustering algorithm. Each class was divided into different clusters to achieve a more accurate prediction.