Conceptual Modeling of Knowledge Bases

Abstract

The task of a knowledge base is to represent a relevant body of knowledge about a specific application domain in order to support knowledge-based application programs and to be able to answer ad-hoc queries about the state of the application domain. It is essential to develop an adequate conceptual model of the application domain as the basis for designing a knowledge base. This involves the formalization of intuitions and of informal descriptions. We discuss the most important methodology for conceptual modeling: entity-relationship modeling which may be viewed as an extension of predicate logic. The emphasis of this chapter is on identifying the different types of predicates needed for capturing more domain knowledge: normal extensional, qualified, incomplete and intensional predicates.