Prédiction des maladies d’insuffisance cardiaque en utilisant des modèles de Deep Learning

Abstract

Heart failure is considered one of the major challenges in the medical field due to its severity and increasing prevalence. In this project, our objective was to develop an artificial intelligence model capable of predicting heart failure cases by combining ECG images with clinical data. To achieve this, we employed Graph Neural Networks (GNN), particularly the GCN and GAT models. Patients and their data were represented as a graph structure to effectively exploit the relationships between them. The experimental results showed that the GCN model delivered better performance compared to the other models used. This work highlights the effectiveness of integrating visual and clinical medical data within a graph-based approach and paves the way for broader adoption of these techniques in medical decision support systems.