Methodology for Dynamic Extraction of Highly Relevant Information Describing Particular Object from Semantic Web Knowledge Base

Exploration and information discovery in a big knowledge base that uses a complex ontology is often difficult, because relevant information may be spread over a number of related objects amongst many other, loosely connected ones. This paper introduces 3 types of relations between classes in an ontology and defines the term of RDF Unit to group relevant and closely connected information. The type of relation is chosen based on association strength in the context of particular ontology. This approach was designed and implemented to manipulate and browse data in a cultural heritage Knowledge Base with over 500M triples, created by PSNC during the SYNAT research project.