Predicting adverse diabetes events in hospitals using machine learning

Project Level: Honours, PhD

We use longitudinal aggregated, routinely collected data from Queensland digital hospitals to develop prediction algorithms for adverse diabetes events based on modern machine learning methods. This will be tested on the historical data from these hospitals and, in the next phase of the project, will be implemented in one or more Queensland hospitals. The projects will also include detailed feature importance and Error analysis of the algorithms developed. The algorithms will help clinicians identify high-risk patients and will lead to improved patient outcomes.