AUG 7, 2026 · PREPRINT
When GNNs Fail: Quantifying and Overcoming Temporal Correlation Volatility in Time Series
arXiv
This is an unrefereed arXiv preprint proposing a novel GNN architecture with empirical validation on synthetic and real-world benchmarks, but lacking peer review and clinical or patient-level outcomes.
Reported
Average performance improvement (…up to 45.6%
Maximum performance gain (GLIDE v…85.7%