pypomp.functional.panel_pfilter¶

pypomp.functional.panel_pfilter(struct: PanelPompStruct, thetas_array: Array, J: int, keys: Array, thresh: float = 0.0, chunk_size: int = 1, CLL: bool = False, ESS: bool = False, filter_mean: bool = False, prediction_mean: bool = False) → dict[str, Array][source]¶

Evaluate panel POMP log-likelihood via particle filtering.

A pure functional implementation of the panel particle filter, intended for composition within custom JAX loops.

Parameters:

struct (PanelPompStruct) – Compiled structural representation of the Panel POMP model.
thetas_array (jax.Array) – Swarm of parameters of shape (n_reps, U, n_params) on the natural scale, aligned with the canonical order of struct.shared_param_names and struct.unit_param_names per unit.
J (int) – Number of particles.
keys (jax.Array) – Random keys of shape (n_reps, U_padded, ...).
thresh (float, optional) – Resampling threshold. Defaults to 0.0.
chunk_size (int, optional) – Number of units to process per chunk. Defaults to 1.
CLL (bool, optional) – Whether to compute conditional log-likelihoods. Defaults to False.
ESS (bool, optional) – Whether to compute effective sample sizes. Defaults to False.
filter_mean (bool, optional) – Whether to compute filtered state means. Defaults to False.
prediction_mean (bool, optional) – Whether to compute prediction state means. Defaults to False.

Returns:

A dictionary containing the results of the panel particle filter. Always contains 'logLik'. Optionally contains 'CLL', 'ESS', 'filter_mean', and 'prediction_mean' if their corresponding flags are True.

Return type:

dict of str to jax.Array

Notes

To align and stack input parameter arrays into the correct canonical ordering, use pypomp.functional.align_params().