Outline of a knowledge-base model for an intelligent information retrieval system

摘要

We attempt in this article to outline an approach to the automatic construction of a knowledge base. We propose some methods and a domain knowledge model. A new idea is to design a system that is able in each phase of its construction to acquire domain knowledge from all new information that it is building, in particular the index terms. We give some methods using production rules to integrate index terms into the domain knowledge. We propose an outline of a formalism for the organization of the different concepts into this knowledge base by means of a semantic network by defining a set of properties. Particular attention is given to the method used to build a very useful property like “synonym”; we give some ideas and experiments for building this relation. In the last section, we describe a user interface to enter rules to be used later for automatic domain knowledge building.