Components: explain#

autogluon.eda.visualization.explain#

`ExplainForcePlot`	Visualize the given SHAP values with an additive force layout
`ExplainWaterfallPlot`	Visualize the given SHAP values with a waterfall layout

ExplainForcePlot#

class autogluon.eda.visualization.explain.ExplainForcePlot(display_rows: bool = False, namespace: Optional[str] = None, **kwargs)[source]#

Visualize the given SHAP values with an additive force layout

Parameters

display_rows (bool, default = False) – if True then display the row before the explanation chart
headers (bool, default = False) – if True then render headers
namespace (str, default = None) – namespace to use; can be nested like ns_a.ns_b.ns_c

Examples

>>> import autogluon.eda.analysis as eda
>>> import autogluon.eda.visualization as viz
>>> import autogluon.eda.auto as auto
>>>
>>> rows_to_explain = ...  # DataFrame
>>>
>>> auto.analyze(
>>>     train_data=..., model=...,
>>>     anlz_facets=[
>>>         eda.explain.ShapAnalysis(rows),
>>>     ],
>>>     viz_facets=[
>>>         viz.explain.ExplainForcePlot(text_rotation=45, matplotlib=True),  # defaults used if not specified
>>>     ]
>>> )

ExplainWaterfallPlot#

class autogluon.eda.visualization.explain.ExplainWaterfallPlot(display_rows: bool = False, namespace: Optional[str] = None, **kwargs)[source]#

Visualize the given SHAP values with a waterfall layout

Parameters

display_rows (bool, default = False) – if True then display the row before the explanation chart
headers (bool, default = False) – if True then render headers
namespace (str, default = None) – namespace to use; can be nested like ns_a.ns_b.ns_c

Examples

>>> import autogluon.eda.analysis as eda
>>> import autogluon.eda.visualization as viz
>>> import autogluon.eda.auto as auto
>>>
>>> rows_to_explain = ...  # DataFrame
>>>
>>> auto.analyze(
>>>     train_data=..., model=...,
>>>     anlz_facets=[
>>>         eda.explain.ShapAnalysis(rows_to_explain),
>>>     ],
>>>     viz_facets=[
>>>         viz.explain.ExplainWaterfallPlot(),
>>>     ]
>>> )

autogluon.eda.analysis.explain#

ShapAnalysis

Perform Shapley values calculation using shap package for the given rows.

ShapAnalysis#

class autogluon.eda.analysis.explain.ShapAnalysis(rows: DataFrame, baseline_sample: int = 100, parent: Optional[AbstractAnalysis] = None, children: Optional[List[AbstractAnalysis]] = None, state: Optional[AnalysisState] = None, random_state: int = 0, **kwargs)[source]#

Perform Shapley values calculation using shap package for the given rows.

Parameters

rows (pd.DataFrame,) – rows to explain
baseline_sample (int, default = 100) – The background dataset size to use for integrating out features. To determine the impact of a feature, that feature is set to “missing” and the change in the model output is observed.
parent (Optional[AbstractAnalysis], default = None) – parent Analysis
children (List[AbstractAnalysis], default = []) – wrapped analyses; these will receive sampled args during fit call
state (AnalysisState) – state to be updated by this fit function
random_state (int, default = 0) – random state for sampling
kwargs –

Examples

>>> import autogluon.eda.analysis as eda
>>> import autogluon.eda.visualization as viz
>>> import autogluon.eda.auto as auto
>>>
>>> auto.analyze(
>>>     train_data=..., model=...,
>>>     anlz_facets=[
>>>         eda.explain.ShapAnalysis(rows, baseline_sample=200),
>>>     ],
>>>     viz_facets=[
>>>         # Visualize the given SHAP values with an additive force layout
>>>         viz.explain.ExplainForcePlot(),
>>>         # Visualize the given SHAP values with a waterfall layout
>>>         viz.explain.ExplainWaterfallPlot(),
>>>     ]
>>> )