Skip to main content

Integration of Ontologies in the Platform AI

Important Notes:​

  1. Ontologies are always bound to particular context.
  2. Ontologies are always bound to particular schema,
  3. schema represents the context.

Note:​

For now the schemas are pre-defined, so the list of pre-defined schemas is limited and for each schema system should register Ontology converter.

As you can see the use cases for using ontologies could be very different depending on context and event which triggers creation of ontologies. The events itself are categorized.

  1. System state change event, when new predictions were stored as an instances, individuals of particular concepts. So, the system should have predefined rules, instance filters(search criteria), for matching already existing instances(individual of corresponding concept), predefined schema. The combination of this 3 basic components rules,instance filters,schema we will call Ontology creation template.

  2. User request, when the user manually creates an instance filters(search criteria), for matching already existing instances(individual of corresponding concept) and requests an ontology creation. In this case we need to have predefined schema to be selected by user.

As you see, in both cases we will have Ontology creation request, based on which we will process ontology creation. We also will create Ontology record, by referencing to Ontology creation request. Based on use case, the reference of this Ontology will be mapped with particular domain specific components for further usage.