Automatic Collocational Retrieval of NYMs
This system identifies semantically-related pairs of words (or 'nyms'), based on the similarity of their collocational environments. The results are being used as a basis for a description of the textual thesaurus, and in document retrieval as an alternative search term facility, and as an automatic thesaurus generator.
The ACRONYM thesaural facility identifies the conventional sense relations of synonymy (e.g. doctor and medic), antonymy (e.g. luxury and no-frills), hyponymy (e.g. doctor and paediatrician), meronymy (e.g. car and engine), and less well-established semantic relations characteristic of textual usage. The lexical realisations of the sense relations are very often different from those found in a representation of the mental lexicon such as Roget's Thesaurus, but they reflect the thesaurus as it is used in text. From a document retrieval point of view, meanwhile, the fact that the word pairs come from similar contexts is as important as the fact that they are semantically related.
The full ACRONYM system identifies single-word and multi-word nyms, disambiguates the nymic output for polysemous words, and identifies change in semantic pairings across time. It also runs on French text and is adaptable to work on and across most languages.
We also have an ACRONYM demo.
Acknowledgement: The ACRONYM Project was funded by the EPSRC from 1994-1997 (Grant Reference GR/J94754/01).