ENHANCING DATABASE-MANAGEMENT TO KNOWLEDGE-BASE MANAGEMENT - THE ROLE OF INFORMATION-RETRIEVAL TECHNOLOGY

In recent years, there has been an enhancement from database management to knowledge base management. In this article, we investigate an important, yet not widely addressed issue: the role of information retrieval in this enhancement. Central to our investigation is to develop techniques of building information systems which are able to extract and store information from texts so that the stored information can be combined to satisfy unanticipated queries. Success of such techniques will have important practical impact on the automatic construction of knowledge bases from documents, thus reducing the knowledge acquisition bottleneck. We argue that this is not a task which can simply be reduced to database management through logic-based approaches (such as deductive databases), and explore an approach to achieve this by extending classical information retrieval techniques. An experiment has been carried out to implement a conceptual model for storage and retrieval of integrating short scientific documents (written in restricted English) using unstructured relational databases. This article does not address the issues dealing with general information retrieval; instead, by restricting our focus on the problem investigated, we have gained some insight on the relationship between information retrieval and database management, as well as the relationship between information retrieval and knowledge base management.