Representing Natural Language Sentences in RDF Graphs to Derive Knowledge Patterns

摘要

English language has its own word order, which determines the way of ordering words in a sentence. This paper presents the manner by which sentences in English language can be represented as knowledge patterns by means of RDF graphs. Knowledge patterns are general patterns of knowledge that can be used in any knowledge base or ontology. Knowledge pattern is a logical design pattern which defines a formal expression that can be exemplified and morphed in order to solve a domain modeling problem. Knowledge patterns are proposed by ontology engineers. When using them, the general symbols in the pattern are renamed to special symbols from the modeled domain. The RDF graph language is used to represent knowledge pattern contained English sentences. The basic RDF model can be processed even in the absence of detailed information (schema) on the semantics. RDF model has the important property of being modular, i.e., the union of knowledge represented as directed graphs. The paper also provides examples of usage of knowledge patterns.