Premium
Automatically building a knowledge base through natural language text analysis
Author(s) -
Hodges Julia E.,
Cordova Jose L.
Publication year - 1993
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.4550080903
Subject(s) - computer science , knowledge base , natural language processing , schema (genetic algorithms) , knowledge based systems , artificial intelligence , domain knowledge , knowledge representation and reasoning , open knowledge base connectivity , natural language , knowledge acquisition , component (thermodynamics) , question answering , domain (mathematical analysis) , expert system , procedural knowledge , information retrieval , knowledge management , personal knowledge management , mathematical analysis , organizational learning , physics , mathematics , thermodynamics
The knowledge representation and acquisition system described in this article is one of two major components in a project intended to provide expert systems with the ability to build and update their own knowledge bases by processing natural language technical material that is in machine‐readable form. the particular aspect of the system described in this article is the building of an initial knowledge base using only syntactic structures (which include semantic information about domain‐dependent object classes) provided by the natural language processing component and a schema‐like description of the knowledge base. Descriptions of the knowledge structures that make up the knowledge base and the approach used to extract the appropriate information from the syntactic structures are provided. the approach taken during the development of the knowledge representation and acquisition component was motivated by two key factors: keeping the algorithms as general (i.e., domain independent) as possible, and making the process as fully automated as possible. the little human intervention that has been used has been limited to certain stages of the natural language processing. © 1993 John Wiley & Sons, Inc.