Prosthetic joint infection (PJI) represents one of the most common reasons for failure among hip and knee arthroplasty, with an incidence of around 1-2%. Infection can occur early (within days of surgery) or late (over a year after surgery), and no specific early markers for infection onset exist. Given the significant costs to the NHS for corrective revision surgery, the added suffering and risks to patients from surgery, and the risk of enhancing antimicrobial resistance through the use of broad-spectrum antibiotics, a more specific predictive test for early onset of infection is required.
Use of TensorFlow through the Keras API in RStudio to explore deep learning model training
Use of multiple machine learning techniques to explore a database of Pokémon, including creation of a recommendation machine and development of prediction alogorithms for determining Legendary Pokémon
An innovative new Research Centre to identify novel enzymatic solutions to environmental waste problems such as plastic
Further exploration of posts from the Ultra Running Community (URC) Facebook page, including using machine learning techniques including logistic regression and random forest to explore the predictibility ofposters based on the content of their posts