Fig. 5: Resolution of clinical uncertainty with InfEHR.
From: InfEHR: Clinical phenotype resolution through deep geometric learning on electronic health records
a Schematic showing key architectural details of the InfEHR Graph Neural Network (GNN) and information flow through the architecture. b InfEHR’s performance on false negatives predicted by the clinical heuristic (top) and on training and internal validation in neonatal culture-negative sepsis. c. Density distribution of predicted probabilities from the InfEHR GNN on patients with and without disease.
