Enrichment Ontology via Linked Data

Ontology Learning

• •

  Ontology learning is the process of acquiring (constructing or integrating) an ontology.

• •

  (Semi-) automatically. The acquisition of ontologies can be performed through three.

• •

  Major approaches: (Petasis, Karkaletsis, Paliouras, Krithara, & Zavitsanos, 2011).

• •

  By integrating existing ontologies. The integration process tries to capture commonalities among ontologies that convey the same or similar domains, in order to derive a new ontology.

• •

  By constructing an ontology from scratch or by extending (enriching) an existing.

• •

  Ontology, usually based on information extracted from domain-specific content.

• •

  By specializing a generic ontology, in order to adapt it to a specific domain. Ontology learning is not simply a replication of existing work under a different name.

• •

  It adds novel aspects to the problem of knowledge acquisition:

• •

  Ontology learning combines research from knowledge representation, logic, philosophy, databases, machine learning, natural language processing, image/audio/video analysis, etc.

• •

  Ontology learning in the context of the Semantic Web must deal with the massive and heterogeneous data of the World Wide Web and thus improve existing approaches for knowledge acquisition, which target mostly small and homogeneous data collections. (Petasis, Karkaletsis, Paliouras, Krithara, & Zavitsanos, 2011)

• •

  Substantial effort is being put into the development of extensive and rigorous evaluation methods in order to evaluate ontology learning approaches on welldefined tasks with well-defined evaluation criteria.

Ontology Enrichment

Ontology enrichment is the process of extending an ontology, through the addition of new concepts, relations and rules. It is performed every time that the existing domain knowledge is not sufficient to explain the information extracted from the corpus or linked data in our case.