Knowledge Extractor: A Tool for Extracting Knowledge from Text

This paper describes a knowledge-base development tool being developed at Virginia Tech to assist domain experts (knowledge engineers) construct domain-specific knowledge bases. The methodology assumed here for knowledge-base development is to begin with a relatively small, basic ontology and extract conceptual graphs from text in the domain of interest. During this process, the ontology is augmented by the knowledge engineer as needed. The user scans the text, constructing conceptual graphs from sentences or other expressions, and joins the individual graphs into a knowledge-base.