MMMSummaryFactory.posterior_predictive#

MMMSummaryFactory.posterior_predictive(hdi_probs=None, frequency=None, output_format=None)[fuente]#

Create posterior predictive summary DataFrame.

Computes mean, median, and HDI bounds for posterior predictive samples, along with observed values for comparison.

Parameters:
hdi_probssequence of float, optional

HDI probability levels (default: uses factory default)

frequency{«original», «weekly», «monthly», «quarterly», «yearly», «all_time»}, optional

Time aggregation period (default: None, no aggregation)

output_format{«pandas», «polars»}, optional

Output DataFrame format (default: uses factory default)

Returns:
pd.DataFrame or pl.DataFrame

Summary DataFrame with columns:

  • date: Time index

  • mean: Mean of posterior predictive samples

  • median: Median of posterior predictive samples

  • observed: Observed target values

  • abs_error_{prob}_lower: HDI lower bound for each prob

  • abs_error_{prob}_upper: HDI upper bound for each prob

Examples

>>> df = mmm.summary.posterior_predictive()
>>> df = mmm.summary.posterior_predictive(frequency="monthly")
>>> df = mmm.summary.posterior_predictive(hdi_probs=[0.80, 0.94])