My Dissertation Research Project in partial fulfilment of my Masters Degree in Computer Science.

Title: 'Federated Learning in Smart Healthcare: Applying Personalisation to Image-Classification of Cardiovascular-Disease'.

This paper focuses on Federated Learning and its usage in the Smart Healthcare domain.

The paper elaborates prevalence of non-IID data trends in healthcare data and the effect it has on machine learning model prediction perfromance.

An approach to combat these effects is developed and tested in extension of an existing federated learning tool that predicts cardiovascular-disease. This approach involves the use of synthetic data augmentation and transfer learning to produce personalised models for clients in a federated learning system.

The personalised models display an 11.31% prediction accuracy increase in comparison to a global model produced by the original classifier.