midr-package: midr: Learning from Black-Box Models by Maximum...

In midr: Learning from Black-Box Models by Maximum Interpretation Decomposition

midr: Learning from Black-Box Models by Maximum Interpretation Decomposition

Description

The goal of 'midr' is to provide a model-agnostic method for interpreting and explaining black-box predictive models by creating a globally interpretable surrogate model. The package implements 'Maximum Interpretation Decomposition' (MID), a functional decomposition technique that finds an optimal additive approximation of the original model. This approximation is achieved by minimizing the squared error between the predictions of the black-box model and the surrogate model. The theoretical foundations of MID are described in Iwasawa & Matsumori (2025) [Forthcoming], and the package itself is detailed in Asashiba et al. (2025) \Sexpr[results=rd]{tools:::Rd_expr_doi("10.48550/arXiv.2506.08338")}.

Author(s)

Maintainer: Ryoichi Asashiba [e-mail]

Authors:

• Hirokazu Iwasawa

Other contributors:

• Reiji Kozuma [contributor]

See Also

Useful links:

• https://github.com/ryo-asashi/midr

• https://ryo-asashi.github.io/midr/

• Report bugs at https://github.com/ryo-asashi/midr/issues

midr documentation built on Sept. 11, 2025, 1:07 a.m.