interpret Package

Contains functionality for working with model interpretability in Azure Machine Learning.

You can use model interpretability to explain why a model makes the predictions it does and help build confidence in the model. With this package you can get feature and class importance for blackbox and whitebox models, on both raw and engineered features. For more information, see the article Model interpretability in Azure Machine Learning.

This package uses the interpretability techniques developed in the Interpret Community SDK, an open source Python package for training interpretable models and helping to explain blackbox systems, with additional interpretability techniques and utility functions to handle real-world datasets and workflows. The Interpret Community SDK hosts the Azure Machine Learning SDK supported explainers such as SHAP explainers, Mimic Explainer, Tabular Explainer, and others.

The key class in this package is the MimicWrapper class, which provides a wrapper to reduce the number of function calls needed to work with the interpret model package.

Packages

common	Contains common infrastructure, class hierarchy, and utilities for model explanations in Azure Machine Learning.
model	Defines model concept for interpretability.
scoring	Module for lightweight explainers to be run at scoring time.

Modules

mimic_wrapper

Defines functionality to wrap machine learning interpretability into a single API.

Classes

ExplanationClient	Defines the client that uploads and downloads explanations. Create the client used to interact with explanations and run history.
MimicWrapper	A wrapper explainer which reduces the number of function calls necessary to use the explain model package. Initialize the MimicWrapper. `<<that accepts a 2d ndarray :param explainable_model: The uninitialized surrogate model used to explain the black box model. Also known as the student model.