Evaluate single emergent constraint

Diagnostic script to evaluate a single emergent constraint.

Description

Establish a single emergent constraint for an arbitrary input variable and an arbitrary target variable. All input datasets need to be one-dimensional and must include a coordinate 'dataset' or 'model' (thus, the data describes a single scalar value for each dataset). All input datasets must be marked with a var_type (either feature, label, prediction_input or prediction_input_error) and a tag, which describes the type of data. This diagnostic supports only a single tag for label and feature. For every tag, a 'reference_dataset' can be specified, which will be automatically considered as prediction_input. If reference_dataset contains '|' (e.g. 'OBS1|OBS2'), multiple datasets are considered as prediction_input (in this case 'OBS1' and 'OBS2').

Author

Manuel Schlund (DLR, Germany)

Project

CRESCENDO

Configuration options in recipe

additional_data: list of dict, optional

Additional datasets given as list of metadata.

all_data_label: str, optional (default: ‘all’)

Label used in plots when all input data is considered. Only relevant if group_by is not used.

combine_groups: bool, optional (default: False)

Add results to plots for data generated by combining the data of all individual groups.

confidence_level: float, optional (default: 0.66)

Confidence level for estimation of target variable.

group_by: str, optional

Group input data by an attribute (e.g. produces separate plots for the individual groups, etc.).

ignore_patterns: list of str, optional

Patterns matched against ancestor files. Those files are ignored.

merge_identical_pred_input: bool, optional (default: True)

Use identical prediction_input values as single value.

numbers_as_markers: bool, optional (default: False)

Use numbers as markers in scatterplots.

patterns: list of str, optional

Patterns matched against ancestor files.

plot_regression_line_mean: bool, optional (default: False)

Plot means of regression lines in scatterplots.

read_external_file: str, optional

Read input datasets from external file given as absolute path or relative path. In the latter case, 'auxiliary_data_dir' from the user configuration file is used as base directory.

savefig_kwargs: dict

Keyword arguments for matplotlib.pyplot.savefig().

seaborn_settings: dict

Options for seaborn.set() (affects all plots).