deepmol.feature_importance package

Submodules

deepmol.feature_importance.shap_values module

class ShapValues(explainer: str = 'permutation', masker: str | None = None)

Bases: object

SHAP (SHapley Additive exPlanations) wrapper for DeepMol It allows to compute and analyze the SHAP values of DeepMol models.

bar_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of all the features as a beeswarm plot.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional keyword arguments for the plot function: see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_bar.py#L19

beeswarm_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of all the features as a beeswarm plot.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional keyword arguments for the plot function: see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_beeswarm.py#L23

decision_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of all the features as a decision plot.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_decision.py#L222

feature_explanation_plot(index: int, path: str | None = None, **kwargs) → None

Plot the SHAP values of a single feature.

Parameters:

• index (int) – Index of the feature to explain

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_scatter.py#L19

fit(dataset: Dataset, model: Model, **kwargs) → Explanation

Compute the SHAP values for the given dataset and model.

Parameters:

• dataset (Dataset) – The dataset to compute the SHAP values for.

• model (Model) – The model to compute the SHAP values for.

• kwargs (dict) – Additional arguments for the SHAP explainer.

Returns:

shap_values – The SHAP values.

Return type:

np.array

heatmap_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of all the features as a heatmap.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_heatmap.py#L12

negative_class_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of the positive class as a bar plot.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_bar.py#L19

positive_class_plot(path: str | None = None, **kwargs) → None

Plot the SHAP values of the positive class as a bar plot.

Parameters:

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see: https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_bar.py#L19

sample_explanation_plot(index: int, plot_type: str = 'waterfall', path: str | None = None, **kwargs) → None

Plot the SHAP values of a single sample.

Parameters:

• index (int) – Index of the sample to explain

• plot_type (str) – Type of plot to use. Can be ‘waterfall’ or ‘force’

• path (str) – Path to save the plot to.

• kwargs – Additional arguments for the plot function. see:https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_force.py#L33 https://github.com/slundberg/shap/blob/45b85c1837283fdaeed7440ec6365a886af4a333/shap/plots/_waterfall.py#L18

Module contents