PanelPomp Class

class pypomp.panel.panel.PanelPomp(Pomp_dict: dict[str, Pomp], theta: PanelParameters | dict[str, DataFrame | None] | list[dict[str, DataFrame | None]] | None = None)

Bases: PanelValidationMixin, PanelEstimationMixin, PanelAnalysisMixin

The PanelPomp class represents a panel of partially observed Markov process models. It extends the single-unit POMP framework to handle multiple units that share structural characteristics but may have distinct parameter values and observations.

In particular, the class provides methods for:

• Simulation of panel models

• Particle filtering for panel models

• Marginalized Panel Iterated Filtering (MPIF)

• Gradient descent via automatic differentiation

Parameters:

• Pomp_dict (dict[str, Pomp]) – A dictionary mapping unit names to Pomp objects. Each Pomp object represents a single unit in the panel data. The keys are used as unit identifiers.

• theta (PanelParameters | dict | list, optional) – A PanelParameters object, a dictionary with “shared” and “unit_specific” keys, or a list of such dictionaries.

Attributes

ys	Returns a tidy DataFrame with the observations from all units.
unit_objects
theta
results_history
fresh_key
metadata
canonical_param_names
canonical_shared_param_names
canonical_unit_param_names

Core Algorithmic Methods

simulate(key[, theta, times, nsim, as_pomp])	Simulate the PanelPomp model.
pfilter(J[, key, theta, thresh, reps, ...])	Run the particle filter (SMC) algorithm on the PanelPomp model.
mif(J, M, rw_sd, a[, key, theta, thresh, ...])	Estimate parameters using the Panel Iterated Filtering (PIF) algorithm for PanelPomp.
train(J, M, eta[, chunk_size, optimizer, ...])	Estimate parameters using chunked gradient-descent optimization (SGD/Adam).

Supporting Methods

get_unit_names()
get_unit_parameters(unit[, theta])
sample_params(param_bounds, units, n, key[, ...])	Sample parameters for PanelPomp models.
prune([n, refill])	Prune the parameter sets to the top n based on log-likelihoods stored in the PanelParameters object under theta.
mix_and_match()	Sorts unit-specific parameters and shared parameters in descending order of unit log-likelihood and shared log-likelihood, respectively, then combines them to form new parameter sets.
results([index, ignore_nan])	Returns a tidy DataFrame with the results of the method run at the given index.
CLL([index, average])	Returns a tidy DataFrame with the conditional log-likelihoods of the method run at the given index.
ESS([index, average])	Returns a tidy DataFrame with the effective sample size of the method run at the given index.
time()	Return a DataFrame summarizing the execution times of methods run.
traces()	Return a tidy DataFrame with the full trace of log-likelihoods and parameters from the entire result history.
probe(probes, key[, nsim, theta])	Evaluate probe statistics on the model's true data and simulated data for each unit.
arma([order, log_ys, suppress_warnings])	Fits an independent ARIMA model to the observation data for each unit and returns a DataFrame with the estimated log-likelihoods for each unit and the total.
negbin([autoregressive, suppress_warnings])	Fits a Negative Binomial model to the observation data for each unit and returns a DataFrame with the estimated log-likelihoods for each unit and the total.
plot_traces([which, show])	Plots the parameter and log-likelihood traces from the entire result history.
plot_simulations(key[, nsim, mode, theta, show])	Runs simulations for the PanelPomp model and plots them against true data.
print_metadata()	Prints the creation and runtime environment metadata for this instance.
print_summary([n])	Print a summary of the PanelPomp object.
merge(*panel_pomp_objs)	Merge replications from multiple PanelPomp objects into a single object.