Causal Interpretations of Black-Box Models

Qingyuan Zhao and Trevor Hastie

Journal of Business & Economic Statistics, 2021, vol. 39, issue 1, 272-281

Abstract: The fields of machine learning and causal inference have developed many concepts, tools, and theory that are potentially useful for each other. Through exploring the possibility of extracting causal interpretations from black-box machine-trained models, we briefly review the languages and concepts in causal inference that may be interesting to machine learning researchers. We start with the curious observation that Friedman’s partial dependence plot has exactly the same formula as Pearl’s back-door adjustment and discuss three requirements to make causal interpretations: a model with good predictive performance, some domain knowledge in the form of a causal diagram and suitable visualization tools. We provide several illustrative examples and find some interesting and potentially causal relations using visualization tools for black-box models.

Date: 2021
References: Add references at CitEc
Citations: View citations in EconPapers (27)

Downloads: (external link)
http://hdl.handle.net/10.1080/07350015.2019.1624293 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:jnlbes:v:39:y:2021:i:1:p:272-281

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/UBES20

DOI: 10.1080/07350015.2019.1624293

Access Statistics for this article

Journal of Business & Economic Statistics is currently edited by Eric Sampson, Rong Chen and Shakeeb Khan

More articles in Journal of Business & Economic Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().