Automatic taxonomy generation: Issues and possibilities

Abstract

Automatic taxonomy generation deals with organizing text documents in terms of an unknown labeled hierarchy. The main issues here are (i) how to identify documents that have similar content, (ii) how to discover the hierarchical structure of the topics and subtopics, and (iii) how to find appropriate labels for each of the topics and subtopics. In this paper, we review several approaches to automatic taxonomy generation to provide an insight into the issues involved. We also describe how fuzzy hierarchies can overcome some of the problems associated with traditional crisp taxonomies.