Sentiment Classification of Movie Reviews: Exploring the use of Neural Networks on Sparse Matrix Encoding

Exploring CNNs on Sparse encodings to classify sentiment. This is adapted from my Applied Machine Learning mini-project at The University of Edinburgh. This project details the cleaning and exploratory data analysis and compares the Keras neural networks used to classify sentiment. The full...

Deep Learning of Non-Equilibrium States

Master's research project for degree, due to be presented in early 2023. This project explores the applications of CNNs and unsupervised clustering to accurately identify the percolation probability of a given directed percolation image, thus isolating the critical value. The performance and...

Drawing Classification and Latent Representations

This is part of my Data Analysis and Machine Learning coursework at The University of Edinburgh. In this exercise, a 2D CNN is used to classify images from the 'quickdraw' dataset with a testing accuracy of 86%. The reconstruction of these images through autoencoders is explored and their latent...

British Airways: Virtual Experience Programme

Task 1: Exploratory data analysis of web-scraped reviews using Beautiful Soup, assigning sentiment with Scikit-Learn. Task 2: Predictive modelling of customer booking completion using a K-fold cross-validated ensemble of decision trees with tuned hyperparameters.