Mining domain ontological information from online publications

Abstract

The authors describe a system that uses online citation information to build reliable domain ontologies. The WWW has made it possible to transfer research literature repositories from hard copies to online databases, but currently there are few computational tools that support scholars\' efforts to interpret and analyze bodies of prior research. The standard process still consists of wading through numerous papers to discover the main threads of a topic, then creating road maps via iterative cycles of identifying themes and other connections among papers. Our goal is to use citation analysis and keyword methods to exploit ontological information found in online scientific literature to establish introductory surveys for researchers entering new fields.