Collocation Mining: Exploiting Corpora for Collocation, Identification and Representation
Brigitte Krenn · 2000
The work presented provides computational linguistics methods and tools for collocation identification from arbitrary text, and methods and tools for representing collocations in a relational database integrating competence (collocation-type-specific linguistic analysis) and performance information (corpus sentences). The work differs from existing approaches to collocation identification in systematically utilizing collocation type-specific linguistic information. With respect to collocation representation, the work is the first to systematically and in a large scale combining competence-based descriptions of collocations with actual occurrences in text. 1