Automatic extraction of terminology under CRF model

Abstract

An automatic terminology extraction method in specific domain is proposed based on condition random fields (CRF) in this paper. We treat extraction of terminology in one domain as a sequence labeling problem, and terminology distribution characteristics as features of the CRF model. Then we use the CRF model to train a template for the terminology extraction. Experimental results show that the method is effective and efficient with common domains. © 2012 Springer-Verlag GmbH.