Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Exploring Ranked Local Selectors for Stable Explanations of ML Models
Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science (from 2013).
Karlstad University, Faculty of Health, Science and Technology (starting 2013), Department of Mathematics and Computer Science (from 2013).ORCID iD: 0000-0003-3461-7079
2021 (English)In: 2021 2nd International Conference on Intelligent Data Science Technologies and Applications, IDSTA 2021, Institute of Electrical and Electronics Engineers (IEEE), 2021, p. 122-129Conference paper, Published paper (Refereed)
Abstract [en]

While complex machine learning methods can achieve great performance, human-interpretable details of their internal reasoning is to a large extent unavailable. Interpretable machine learning can remedy the lack of access to model reasoning but remains an elusive feat to fully achieve. Here we propose ranked selectors as a method for post-hoc explainability of classification outcomes from arbitrary classification models, with an initial emphasis on tabular data of moderate dimensions. The method is based on constructing a set of selectors, or rules, delimiting a local class consistent domain with maximal cover around the item of interest. The extended adjacent feature space is probed to achieve a ranking of the selectors. The method supports the use of an explicit foil class Q, allowing the formulation of contrastive queries in the form 'Why inference P instead of alternative inference Q?'. The answer is given as a short list of disjoint rules, a format previously shown to be amenable to human interpretation. We demonstrate the proposed method in open datasets, and elaborate on its stability aspects relative to other comparable methods.

Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2021. p. 122-129
Keywords [en]
black box models, contrastivity, Explainability, Adjacent feature, Black box modelling, Classification models, Machine learning methods, Maximal covers, Model reasonings, Performance, Tabular data, Machine learning
National Category
Information Studies Computer Sciences
Identifiers
URN: urn:nbn:se:kau:diva-89479DOI: 10.1109/IDSTA53674.2021.9660809ISI: 000852877600018Scopus ID: 2-s2.0-85124559014ISBN: 9781665421805 (print)OAI: oai:DiVA.org:kau-89479DiVA, id: diva2:1651497
Conference
2nd International Conference on Intelligent Data Science Technologies and Applications, IDSTA 2021, 15 November 2021 through 16 November 2021
Available from: 2022-04-12 Created: 2022-04-12 Last updated: 2022-09-22Bibliographically approved

Open Access in DiVA

No full text in DiVA

Other links

Publisher's full textScopus

Authority records

Korhonen, TopiGarcia, Johan

Search in DiVA

By author/editor
Korhonen, TopiGarcia, Johan
By organisation
Department of Mathematics and Computer Science (from 2013)
Information StudiesComputer Sciences

Search outside of DiVA

GoogleGoogle Scholar

doi
isbn
urn-nbn

Altmetric score

doi
isbn
urn-nbn
Total: 140 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • apa.csl
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf