B-SPARQL: A typed language for Querying the Big Knowledge

Abstract

We introduce B-SPARQL as an extension of the classical SPARQL to deal with the problem of querying big knowledge. B-SPARQL extends SPARQL mainly in two directions. The first one is to query big knowledge in a global way such as `how much of the knowledge items of the knowledge base are confidential?’ The second one is to query big knowledge in its translated digital form instead of in its original symbolic form. The former is necessary when evaluating and maintaining an immense knowledge base. The latter is useful when it is very inefficient or even impossible to deal with an immense knowledge base. In accordance, B-SPARQL not only deals with basic knowledge elements such as subjects, predicates and objects of RDF triples. It also deals with triples and knowledge bases and knowledge areas (parts of knowledge bases), including SPARQL endpoints, immediately by considering them as basic knowledge elements as well.