Unsupervised Induction of IE Domain Knowledge using an Ontology

Mark Stevenson

The development of systems which can be easily adapted to new domains is an important goal in current Information Extraction (IE) research. Machine learning algorithms have been applied to the problem but supervised algorithms often require large amounts of examples and unsupervised ones may be hampered by a lack of information. This paper presents an unsupervised algorithm which makes use of the WordNet ontology to compensate for the small number of examples. Comparative evaluation with a previously reported approach shows that the algorithm presented here is in some ways preferable and that benefits can be gained from combining the two approaches.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.