MMMCVPlotSuite.crps

MMMCVPlotSuite.crps(cv_data=None, dims=None, figsize=None, backend=None, return_as_pc=False, line_kwargs=None, **pc_kwargs)

Line chart of mean CRPS per fold for train and test splits.

Renders an n×2 grid: left column = train CRPS, right column = test CRPS, one row per Cartesian combination of extra dimensions in y_original_scale (e.g. one row per geo). When no extra dimensions are present the result is a 1×2 grid.

Parameters:

cv_dataaz.InferenceData or None

Override the stored self.cv_data for this call only.

dimsdict or None

Filters which coordinate values of extra dimensions appear as rows (e.g. {"geo": ["geo_b"]} → only geo_b row). Non-extra-dim keys are silently ignored.

figsizetuple or None

Figure size in inches.

backendstr or None

PlotCollection backend.

return_as_pcbool

Return the raw PlotCollection instead of (Figure, NDArray[Axes]).

line_kwargsdict or None

Extra kwargs forwarded to azp.visuals.line_xy.

**pc_kwargs

Forwarded to PlotCollection.grid().

Returns:

tuple[Figure, NDArray[Axes]] or PlotCollection