Extracting clusters of specialist terms from unstructured text

Abstract

Automatically identifying related specialist terms is a difficult and important task required to understand the lexical structure of language. This paper develops a corpus-based method of extracting coherent clusters of satellite terminology - Terms on the edge of the lexicon - using co-occurrence networks of unstructured text. Term clusters are identified by extracting communities in the cooccurrence graph, after which the largest is discarded and the remaining words are ranked by centrality within a community. The method is tractable on large corpora, requires no document structure and minimal normalization. The results suggest that the model is able to extract coherent groups of satellite terms in corpora with varying size, content and structure. The findings also confirm that language consists of a densely connected core (observed in dictionaries) and systematic, semantically coherent groups of terms at the edges of the lexicon.