pypomp.PanelPomp.arma

PanelPomp.arma(order: tuple[int, int, int] = (1, 0, 1), log_ys: bool = False, suppress_warnings: bool = True) → DataFrame

Fit an independent ARIMA model to the observation data of each unit.

This is a wrapper around pypomp.benchmarks.arma().

Parameters:

  • order (tuple of (int, int, int), optional) – The (p, d, q) order for the ARIMA model. Defaults to (1, 0, 1).

  • log_ys (bool, optional) – If True, fit the model to log(y + 1). Defaults to False.

  • suppress_warnings (bool, optional) – If True, suppress statsmodels warning messages during fitting. Defaults to True.

Returns:

DataFrame with columns 'unit' and 'logLik' containing the individual unit log-likelihoods and their sum (labeled as '[[TOTAL]]').

Return type:

pd.DataFrame