Abstract: The paper presents the results of automatic term extraction from a special text corpus (a collection of papers on corpus linguistics) by means of statistical methods (association measures) combined with certain syntactic models. The approach undertaken in the paper is based on lexico-syntactic models that can be viewed as models of phrases for the Russian language. The Sketch Engine system represents itself a corpus tool which takes as input a corpus of any language and corresponding grammar patterns. The system gives information about a word’s collocability on concrete dependency models, and generates lists of the most frequent phrases for a given word based on appropriate models. The extracted terms belong to various clusters and represent the lexical structure of the texts in question. The applied method includes statistical analysis that enables estimating paradigmatic and syntagmatic relations between lexemes based on their distribution.

Keywords: Corpora, distributional and statistical methods, collocations, automatic term extraction, thesaurus.

ACM Classification Keywords: I.2.7 Natural Language Processing

Link:

STUDYING SPECIAL TEXT RUSSIAN CORPORA BY THE LEXICO-SYNTACTIC MODELS

Maria Khokhlova, Victor Zakharov

http://foibg.com/ibs_isc/ibs-27/ibs-27-p12.pdf