Clinical prediction models and interpretable machine learning
Harry Freitas Da Cruz
We focus on the application of machine learning models to tackle prediction tasks in Nephrology, particularly acute kidney, dialysis outcomes and kidney transplantation using machine learning (ML) methods. ML-based techniques and models have been shown to provide superior predictive accuracy in a number of domains as compared to traditional approaches such as logistic regression, a mainstay in medicine. However, lack of transparency and interpretability in the predictions generated by these complex algorithms has hindered the adoption of ML models in the clinical setting. We aim to 1) apply state of the art in interpretable ML research and 2) develop and test new approaches to make black-box models more interpretable and therefore more accessible to medical practitioners.