MMMCVSummaryFactory

class pymc_marketing.mmm.summary.cv.MMMCVSummaryFactory(cv_data)

Factory for creating summary DataFrames from CV DataTree results.

Obtain via cv.summary on TimeSliceCrossValidator or cv.plot.summary on MMMCVPlotSuite.

Parameters:

cv_dataxr.DataTree

Combined DataTree produced by TimeSliceCrossValidator.run().

Examples

>>> cv_idata = cv.run(X, y, mmm=mmm)
>>> df = cv.summary.predictions()
>>> records = df.to_dict(orient="records")

Methods

MMMCVSummaryFactory.__init__(cv_data)
MMMCVSummaryFactory.crps([dims, output_format])	Create mean CRPS summary per CV fold and split.
MMMCVSummaryFactory.param_stability([...])	Summarize parameter posteriors across CV folds.
MMMCVSummaryFactory.predictions([hdi_probs, ...])	Create posterior predictive summary per CV fold in long format.