Towards semi-automatic extraction of lexical semantics relations for Polish. (26th September 2008)

• Record Type: Journal Article
• Authors: Piasecki, Maciej
• Is Part Of: International journal of intelligent information and database systems. Volume 2:Number 3(2008)
• Page Start: 336
• Page End: 353
• Publication Date: 2008-09-26
• Subjects: semantic similarity -- similarity extraction matrix -- corpus -- WordNet -- nouns -- Poland -- LSA Polish wordnet -- plWordNet -- lexical semantics; Database management -- Computer programs -- Periodicals; Information retrieval -- Computer programs -- Periodicals; Information storage and retrieval systems -- Computer programs -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Intelligent agents (Computer software) -- Periodicals; 006.33
• Languages: English
• ISSNs: 1751-5858

Abstract

A recently started project on building the Polish wordnet (called plWordNet) is the background to the presented work. The goals of the project and the linguistic structure of plWordNet are described. The assumptions of the planned semi-automatic method of extending the initial core wordnet to its full size are presented. A multi-criteria algorithm of extraction of lexical semantic relations is presented. Several problems with application of the known methods of relation extraction to Polish – an inflectional and free word order language, are discussed. Experiments with the application of LSA to Polish news corpus and the extraction of noun similarity on the basis of co-occurrence of nouns and adjectives are presented. The experiments verify the assumed solutions to the problems of large number of Polish word forms, and the lack of sequential structure in Polish.