To address the trade-off between noisy full-graph modeling and information loss in simplified graphs, we propose a Progressive Multi-View Graph Distillation paradigm (PMVGD) for health event ...