RELATIONAL MODELS OF THE LEXICON AND KNOWLEDGE 'REPRESENTATION . : AN OBJECT-ORIENTED APPROACH

Abstract

The Lexicon plays an important role in Natural Language Processing (NLP) and so, an efficient model of the Lexicon is a prerequisite for building Computer systems, which interact gracefully with human beings. Similarly Knowledge Base is a key component in Expert Systems and is responsible for the overall performance of the Expert System. In the first part of the thesis, various Relational models of the Lexicon are discussed and an Object-Oriented model is constructed. The facility for Adding, Deleting, Updating, and Searching of words in such a Lexicon is provided. The search time is considerably reduced by indexing the Lexicon. In the second part of the thesis, different Knowledge representation schemes are discussed at length and finally an Object-Oriented Knowledge Base for Z-Expert (Expert System for configuring personal computers), is constructed. Knowledge Acquisition facility is provided with this Knowledge Base. All the programs for developing the Lexicon and Knowledge base are written in Borland C++, running on an IBM PC/AT.