MMMPlotSuite.plot_cv_predictions

MMMPlotSuite.plot_cv_predictions(results, dims=None)

Plot posterior predictive predictions across CV folds.

Args:

results: List of TimeSliceCrossValidationResult objects returned by run(). dims: Optional dict specifying dimensions to filter when plotting.

Currently only supports {“geo”: […]}. If omitted, all geos present in the posterior predictive y_original_scale are plotted.

The plot shows per-geo, per-fold posterior predictive HDI (3%-97%) for train (blue) and test (orange) ranges as shaded bands, the posterior mean as dashed lines, and observed values as black lines. A vertical dashed green line marks the end of the training period for each fold.