Explainable Hypergraph Neural Networks for the Prediction of Dementia Progression

I conducted my MSc Computer Science dissertation in the field of medical informatics. Specifically, I wanted to explore if it is possible to reliably classify the risk of dementia progression using only clinical data with an emerging Machine Learning technique using hypergraphs to express relationships between risk factors.

The project was succesful and achieved a very good classifiction accuracy of 79% supported by robust capability distinguishing true/false negative and positive cases.

The dissertation achieved a distinction award.

The code is available here.

Video overview

Read the dissertation

You can download the dissertation here or read the dissertation below.