This ontology captures several types of explanation that occur in the literature on explainable machine learning, Bayesian networks and rule-based expert systems. Each explanation consists of an explanandum (that which has to be explained) and an explanans (the part that performs the explaining the explanandum). The explanandum is usually just a fact, or more generally an event. The explanans can take multiple forms: set of facts combined with a rule, a set of important features, a prototype, etc, or a combination of those. (from ontology)
A provider turns a local unique identifiers from a resource into a URI. Many providers are also resolvable as URLs (i.e., they can be used in a web browser).
The local unique identifier isConceptualizedBy is used to demonstrate the providers
available for Explanation Ontology. Some providers may use a different example, which is displayed in the table below.
A guide for curating additional providers can be found
here.
| Provider Name | Provider Code | URI |
|---|---|---|
| Explanation Ontology (RDF) | rdf |
https://www.tno.nl/ontology/knowledgeBaseExplanation#isConceptualizedBy |
| Bioregistry | bioregistry |
https://bioregistry.io/tno.kbe:isConceptualizedBy |