recovery_from_estimates

bayesflow.diagnostics.recovery_from_estimates(estimates: ~collections.abc.Mapping[str, ~collections.abc.Mapping[str, ~numpy.ndarray]], targets: ~collections.abc.Mapping[str, ~numpy.ndarray], marker_mapping: dict[str, str], variable_keys: ~collections.abc.Sequence[str] = None, variable_names: ~collections.abc.Sequence[str] = None, add_corr: bool = True, corr_point_agg: ~collections.abc.Callable = <function median>, figsize: ~collections.abc.Sequence[int] = None, label_fontsize: int = 16, title_fontsize: int = 18, metric_fontsize: int = 16, tick_fontsize: int = 12, color: str = '#132a70', num_col: int = None, num_row: int = None, xlabel: str = 'Ground truth', ylabel: str = 'Estimate', **kwargs) → Figure

Creates and plots publication-ready recovery plot of estimates vs. targets.

This plot yields similar information as the “posterior z-score”, but allows for generic point and uncertainty estimates:

https://betanalpha.github.io/assets/case_studies/principled_bayesian_workflow.html

Important: Posterior aggregates play no special role in Bayesian inference and should only be used heuristically. For instance, in the case of multi-modal posteriors, common point estimates, such as mean, (geometric) median, or maximum a posteriori (MAP) need to be interpreted carefully.

Parameters:

estimatesdict[str, dict[str, np.ndarray]]

The model-generated estimates in a nested dictionary, e.g. as returned by approximator.estimate(conditions=…). - The outer keys identify the inference variable. - The inner keys identify point estimates. - The inner value is an ndarray of shape (num_datasets, point_estimate_size, variable_block_size)

targetsdict[str, np.ndarray]

The prior draws (true parameters) used for generating the num_datasets

marker_mappingdict[str, str]

Define how to mark different point estimates by their key, e.g. {“quantiles”:”_”, “mean”:”*”}. Only point estimates whose key appears in the dictionary, will be plotted.

variable_keyslist or None, optional, default: None

Select keys from the dictionaries provided in estimates and targets. By default, select all keys.

variable_nameslist or None, optional, default: None

The individual parameter names for nice plot titles. Inferred if None

add_corrboolean, default: True

Should correlations between estimates and ground truth values be shown?

corr_point_aggCallable

Function producing a central point estimate from the whole list of point estimates in case correlations should be computed. Default: median

figsizetuple or None, optional, defaultNone

The figure size passed to the matplotlib constructor. Inferred if None.

label_fontsizeint, optional, default: 16

The font size of the y-label text.

title_fontsizeint, optional, default: 18

The font size of the title text.

metric_fontsizeint, optional, default: 16

The font size of the metrics shown as text.

tick_fontsizeint, optional, default: 12

The font size of the axis ticklabels.

colorstr, optional, default: ‘#8f2727’

The color for the true vs. estimated scatter points and error bars.

num_rowint, optional, default: None

The number of rows for the subplots. Dynamically determined if None.

num_colint, optional, default: None

The number of columns for the subplots. Dynamically determined if None.

xlabel:

ylabel:

Returns:

fplt.Figure - the figure instance for optional saving

Raises:

ShapeError

If there is a deviation from the expected shapes of estimates and targets.