Joint Geographic Forecasting

Background: Forecasting multiple geographic locations simultaneously allows models to learn complex inter-state or inter-regional relationships that may improve accuracy, similar to successes seen in hierarchical modeling for influenza.

Question / Future Work: A future direction involves adapting the input/output framework to forecast geographic locations jointly, rather than individually. This requires developing training examples where recent observations from all US states are the input, and the next $H$ time steps for each state are the output, leveraging multi-location capabilities in synthetic data generators like MutAntiGen.