Predicting COVID Patient Risk #

Performed a variety of machine learning techniques on data from COVID hospitalisations in Brazil to develop a model which would be able to accurately predict a patients risk of dying based on their characteristics.

After comparing Logistic Regression, Lasso, Ridge, Polynomial models, and Random Forests, the Random Forest model performed best, achieving strong predictive accuracy.

Feature-importance analysis highlighted age, ICU and hospital stay duration, fever, gender, and vaccination status as key factors associated with severe outcomes.

This project demonstrates how data-driven modelling can uncover clinically relevant patterns and support medical decision-making.

Link to PDF of Report

Technology Utilised: R Studio/R, Microsoft Word, Microsoft Excel