2018

An Interpretable Model with Globally Consistent Explanations for Credit Risk

Chen, Chaofan, Lin, Kangcheng, Rudin, Cynthia et al.

Understand

We propose a possible solution to a public challenge posed by the Fair Isaac Corporation (FICO), which is to provide an explainable model for credit risk assessment.

  • Rather than present a black box model and explain it afterwards, we provide a globally interpretable model that is as accurate as other neural networks.
  • Our "two-layer additive risk model" is decomposable into subscales, where each node in the second layer represents a meaningful subscale, and all of the nonlinearities are transparent.
  • We provide three types of explanations that are simpler than, but consistent with, the global model.

Built on

  • A threshold of ln n for approximating set cover (preliminary version)

    Uriel Feige · 1996

    Earlier work this paper cites.

  • A threshold of ln n for approximating set cover

    Uriel Feige · 1998

    Earlier work this paper cites.

  • Using Data Mining for Bank Direct Marketing: An Application of the CRISP-DM Methodology

    S. Moro, R. Laureano, and P. Cortez · 2011

    Earlier work this paper cites.

  • Accurate intelligible models with pairwise interactions

    Yin Lou, Rich Caruana, Johannes Gehrke, and Giles Hooker · 2013

    Earlier work this paper cites.

  • A data-driven approach to predict the success of bank telemarketing

    Sérgio Moro, Paulo Cortez, and Paulo Rita · 2014

    Earlier work this paper cites.

Similar

  • General data protection regulation

    Protection Regulation · 2016

    Cited alongside, same era.

  • Supersparse linear integer models for optimized medical scoring systems

    Berk Ustun and Cynthia Rudin · 2016

    Cited alongside, same era.

  • Interpretable & explorable approximations of black box models

    Original

    Himabindu Lakkaraju, Ece Kamar, Rich Caruana, and Jure Leskovec · 2017

    Cited alongside, same era.

  • Optimized risk scores

    Berk Ustun and Cynthia Rudin · 2017

    Cited alongside, same era.

  • An optimization approach to learning falling rule lists

    Chaofan Chen and Cynthia Rudin · 2018

    Cited alongside, same era.

  • Local rule-based explanations of black box decision systems

    Original

    Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Dino Pedreschi, Franco Turini, and Fosca Giannotti

    Cited in the paper.

  • A survey of methods for explaining black box models

    Riccardo Guidotti, Anna Monreale, Salvatore Ruggieri, Franco Turini, Fosca Giannotti, and Dino Pedreschi

    Cited in the paper.

Then

  • Mdcalc - medical calculators, equations, algorithms, and scores

    Medical calculators · 2018

    Closest in time.

  • Rulematrix: Visualizing and understanding classifiers with rules

    Y. Ming, H. Qu, and E. Bertini · 2018

    Closest in time.

  • Anchors: High-precision model-agnostic explanations

    Marco Tulio Ribeiro, Sameer Singh, and Carlos Guestrin · 2018

    Closest in time.

  • Globally-consistent rule-based summary-explanations for machine learning models, with application to credit-risk evaluation

    Yaron Shaposhnik and Cynthia Rudin · 2018

    Closest in time.

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…