… In this work, we incorporate a two-layer graph convolutional network (GCN) [9] to encode the
graph representation of clinical notes, as depicted in Figure 1. We include an attention layer …

… clinically to predict hospital readmissions, such as … -risk of being readmitted, particularly
for young patients with rare diseases or the elderly [6], [7]. Furthermore, rule-based or linear risk …

… Specifically, we represent hospital admissions as a graph, where each node in the graph …
scores indicating higher risks of readmission, we treat LACE+ scores as probabilities of …

… for building a good readmission risk prediction model. … readmission prediction with
class-imbalance data, in this paper, we propose a new graph-based prediction method, namely …

… techniques used to evaluate patient readmission risk utilize various structured data in …
Graph Convolutional Network (GCN) [21] for node classification (readmission or not) on a graph …

… Early prediction of readmission risk can assist … graphs to represent discharge summaries.
Then, they trained them with graph convolutional networks to predict ICU readmission rates of …

… ICU readmission risk that uses Knowledge Graph embeddings… Graph Convolutional
Transformers were used to jointly learn the structure of EHR data while performing ICU readmission …

… While current standard of HAI risk prediction utilizes only a narrow set of predefined clinical
variables, we propose a graph convolutional neural network (GNN)-based model which …

… graphs as signals processed through graph convolution in the spectral domain. In particular,
graph … 5, n = 36), followed by mortality and readmission. The diagnosis task also tested the …

… We propose a novel deep learning approach to predict patient’s readmission risk in 30
days. This approach is called TADEL (Trajectory-Aware DEep Learning). It is composed of four …