This dissertation will extend the statistical validation proposed by Miranda et al. which identified the correlation between PhrenicAmp and ALSFRS-R. We will conduct work focused on forecasting the prognosis of ALS using machine learning approaches. It will follow a recent and open research line in the area of Explainable Artificial Intelligence, a research area dedicated to creating AI models that deliver not only their predictions but also, an associated explanation in a sort of human intelligible format.
This work will open the possibilities of using PhrenicAmp as an independent predictor of functional decline in ALS, replacing the FVC which has been used as the gold standard by using machine learning approaches with emphasis on explainability and transparency, facilitating the adoption in real-world clinical scenarios.
This dissertation is aligned with the scientific roadmap of the XAIMED proposal. The student will work in one of the three datasets which were declared at the XAIMED proposal. The student will consolidate the ongoing AICOS’s developments on XAI for time series, propose new approaches and centralise the developments towards the creating of an XAI component for time series.
Author: Maria Antunes
Type: MSc thesis
Partner: FCT NOVA – Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa