The framework consists of three levels:
- knowledge base: It uses Jena TDB as triple store for storing meta-information of ontologies (i..e namespace and URL), list of the registered data and models web services, and thesaurus of standard vocabularies. It can create a URIs for the elements that are not serviced. It stores these types of information under four main concepts space, time, standard vocabularies, and resources
- Knowledge management: It contains four functional components
- Resources harvester: it is context-based recommendation system that works with models and data
- for a model can propose workflows based on the I/O
- for a dataset can provide its contextual relationships with other data
- can suggest models in based on the available data collection
- Logic ingestion:
- ingests new standard vocabularies thesaurus and infers their relationships based on schema matching and string matching
- Linked vocabularies network
- Reasoner:
- validates and extracts facts from the ontologies stored in the knowledge-base, we started by using Pellet reasoner but we plan to add more reasoners such as KAON2
- validates and extracts facts from the ontologies stored in the knowledge-base, we started by using Pellet reasoner but we plan to add more reasoners such as KAON2
- Semantic processor:
- SPARQL query
- Provide semantic mediation based on SKOS standards
- Semantic alignment between resources
- Composition of RDF tags
- Resources harvester: it is context-based recommendation system that works with models and data
- Web services:
- Knowledge discovery:
- Registers resources and their meta-information
- Takes search statements from external resources
- Semantic tagging:
- Retrieves the URL of an ontology or standard name thesauri and can manipulate the elements of an ontology (add, edit, delete)
- Collects tags from registered resources (e.g. SEAD, DataOne)
- Data alignment:
- checks the consistency of the exchanged items between a model and another model or data
- checks the consistency of the exchanged items between a model and another model or data
- Ontology mapping:
- Collects Vocabularies from the Semantic Mediawiki and match them with the registered ontologies
- Recommends relationships between controlled vocabularies
- Knowledge discovery: