shapiq ¶

TabPFN explainer adapters that use the ShapIQ library for model interpretability.

This module provides functions to create shapiq explainers for TabPFN models that support both basic Shapley values and interaction indices for more detailed model explanations.

get_tabpfn_explainer ¶

get_tabpfn_explainer(
    model: TabPFNRegressor | TabPFNClassifier,
    data: DataFrame | ndarray,
    labels: DataFrame | ndarray,
    index: str = "k-SII",
    max_order: int = 2,
    class_index: int | None = None,
    **kwargs
)

Get a TabPFNExplainer from shapiq.

This function returns the TabPFN explainer from the shapiq[1]_ library. The explainer uses a remove-and-recontextualize paradigm of model explanation[2][3] to explain the predictions of a TabPFN model. See shapiq.TabPFNExplainer documentation for more information regarding the explainer object.

Parameters:

Name	Type	Description	Default
`model`	`TabPFNRegressor or TabPFNClassifier`	The TabPFN model to explain.	required
`data`	`DataFrame or ndarray`	The background data to use for the explainer.	required
`labels`	`DataFrame or ndarray`	The labels for the background data.	required
`index`	`str`	The index to use for the explanation. See shapiq documentation for more information and an up-to-date list of available indices. Defaults to "k-SII" and "SV" (Shapley Values like SHAP) with `max_order=1`.	`'k-SII'`
`max_order`	`int`	The maximum order of interactions to consider. Defaults to 2.	`2`
`class_index`	`int`	The class index of the model to explain. If not provided, the class index will be set to 1 per default for classification models. This argument is ignored for regression models. Defaults to None.	`None`
`**kwargs`		Additional keyword arguments to pass to the explainer.	`{}`

Returns:

Type	Description
	shapiq.TabPFNExplainer: The TabPFN explainer.

References

.. [1] shapiq repository: https://github.com/mmschlk/shapiq .. [2] Muschalik, M., Baniecki, H., Fumagalli, F., Kolpaczki, P., Hammer, B., Hüllermeier, E. (2024). shapiq: Shapley Interactions for Machine Learning. In: The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track. url: https://openreview.net/forum?id=knxGmi6SJi .. [3] Rundel, D., Kobialka, J., von Crailsheim, C., Feurer, M., Nagler, T., Rügamer, D. (2024). Interpretable Machine Learning for TabPFN. In: Longo, L., Lapuschkin, S., Seifert, C. (eds) Explainable Artificial Intelligence. xAI 2024. Communications in Computer and Information Science, vol 2154. Springer, Cham. https://doi.org/10.1007/978-3-031-63797-1_23

get_tabpfn_imputation_explainer ¶

get_tabpfn_imputation_explainer(
    model: TabPFNRegressor | TabPFNClassifier,
    data: DataFrame | ndarray,
    index: str = "k-SII",
    max_order: int = 2,
    imputer: str = "marginal",
    class_index: int | None = None,
    **kwargs
)

Gets a TabularExplainer from shapiq with using imputation.

This function returns the TabularExplainer from the shapiq[1][2] library. The explainer uses an imputation-based paradigm of feature removal for the explanations similar to SHAP[3]_. See shapiq.TabularExplainer documentation for more information regarding the explainer object.

Parameters:

Name	Type	Description	Default
`model`	`TabPFNRegressor or TabPFNClassifier`	The TabPFN model to explain.	required
`data`	`DataFrame or ndarray`	The background data to use for the explainer.	required
`index`	`str`	The index to use for the explanation. See shapiq documentation for more information and an up-to-date list of available indices. Defaults to "k-SII" and "SV" (Shapley Values like SHAP) with `max_order=1`.	`'k-SII'`
`max_order`	`int`	The maximum order of interactions to consider. Defaults to 2.	`2`
`imputer`	`str`	The imputation method to use. See `shapiq.TabularExplainer` documentation for more information and an up-to-date list of available imputation methods.	`'marginal'`
`class_index`	`int`	The class index of the model to explain. If not provided, the class index will be set to 1 per default for classification models. This argument is ignored for regression models. Defaults to None.	`None`
`**kwargs`		Additional keyword arguments to pass to the explainer.	`{}`

Returns:

Type	Description
	shapiq.TabularExplainer: The TabularExplainer.

References

.. [1] shapiq repository: https://github.com/mmschlk/shapiq .. [2] Muschalik, M., Baniecki, H., Fumagalli, F., Kolpaczki, P., Hammer, B., Hüllermeier, E. (2024). shapiq: Shapley Interactions for Machine Learning. In: The Thirty-eight Conference on Neural Information Processing Systems Datasets and Benchmarks Track. url: https://openreview.net/forum?id=knxGmi6SJi .. [3] Lundberg, S. M., & Lee, S. I. (2017). A Unified Approach to Interpreting Model Predictions. Advances in Neural Information Processing Systems 30 (pp. 4765--4774).