Add optional time variation directly to all ascertainment models. Each model samples scalar baseline rates—independent, jointly distributed, or ratio-linked—then applies an optional signal-specific temporal process on the logit scale. Signals without a temporal process remain fixed.
Time-varying ascertainment applies on the observation date, after delay convolution, and broadcasts across subpopulations. Centralize validation and NumPyro naming in the base class, add IndependentAscertainment.
Add optional time variation directly to all ascertainment models. Each model samples scalar baseline rates—independent, jointly distributed, or ratio-linked—then applies an optional signal-specific temporal process on the logit scale. Signals without a temporal process remain fixed.
Time-varying ascertainment applies on the observation date, after delay convolution, and broadcasts across subpopulations. Centralize validation and NumPyro naming in the base class, add
IndependentAscertainment.