Explainability in RIONA Algorithm Combining Rule Induction and Instance-Based Learning

• Autorzy: Góra Grzegorz , Skowron Andrzej , Wojna Arkadiusz

Identyfikatory

DOI

10.15439/2023F4139

• Języki publikacji: EN

Abstrakty

EN

The article concerns the well-known RIONA algorithm. We focus on the explainability property of this algorithm. The theoretical results, formulated and proved in the paper, show the relationships of the RIONA classifiers to both instance- and rule-based classifiers. In particular, we show the equivalence (relative to the classification) of the RIONA algorithm with the rule-based algorithm generating all consistent and maximally general rules from the neighbourhood of the test case. Consequently, the RIONA classifier can be represented by a rule-based classifier, with rules easily interpretable by humans. These theoretical results provide the explainability of the classifiers generated by RIONA and could be used in situations when an explanation or justification of the derived decision is important. It should be noted that the RIONA algorithm requires analysing only a small number of objects and rules contrary to algorithms based on the generation of huge sets of rules.

• Wydawca: Polskie Towarzystwo Informatyczne
• Czasopismo: Annals of Computer Science and Information Systems
• Rocznik: 2023
• Tom: Vol. 35
• Strony: 491--502
• Opis fizyczny: Bibliogr. 35 poz.

Twórcy

autor

Góra Grzegorz

autor

Skowron Andrzej

autor

Wojna Arkadiusz

Uwagi

1. Thematic Tracks Regular Papers

2. Opracowanie rekordu ze środków MEiN, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2024).