pypomp.Pomp.CLL

Pomp.CLL(index: int = -1, average: bool = False) → DataFrame

Return conditional log-likelihoods from a particle filter run.

Parameters:

• index (int, optional) – Position in results_history to retrieve. Defaults to -1 (the most recent run). The indexed result must be a pfilter() result with CLL=True.

• average (bool, optional) – If True, average the CLLs over replicates using logmeanexp. Defaults to False.

Returns:

Tidy DataFrame of conditional log-likelihoods. The columns appear in the following order:

1. theta_idx: The index of the parameter set.

2. rep: The replicate index (only if average=False).

3. time: The observation time point.

4. CLL: The conditional log-likelihood value.

Return type:

pd.DataFrame