Wyniki wyszukiwania - BazTech

Ograniczanie wyników

1 ECONTECHMOD : An International Quarterly Journal on Economics of Technology and Modelling Processes

1 2017

Znaleziono wyników: 1

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: semantic parsing

Sortuj według:

Ogranicz wyniki do:

Partial semantic parsing of sentences by means of grammatically augmented ontology and weighted affix context-free grammar

Davydov M., Lozynska O., Pasichnyk V.

ECONTECHMOD : An International Quarterly Journal on Economics of Technology and Modelling Processes

2017

Vol. 6, No 2

27--32

In spite of the fact that modern statistical and neural net based tools for parsing natural language texts supersede classical approaches there are still areas where generative grammars are used. These are areas where collection of universal parallel corpuses is still in the progress. National sign languages are among them. Ontologies and common sense databases play valuable role in parsing and translation of such languages. Grammatically augmented ontology (GAO) is an ontology extension that links phrases to their meaning. The link is established via special expressions that connect phrase meaning to grammatical and semantical attributes of words that constitute it. The article introduces a new approach to sentence parsing that is based on integration of ontology relations into productions of weighted affix context-free grammar (WACFG). For that reason a new parser for WACFG grammar was developed inspired by works of C.H.A. Koster. Basic properties of WACFG are discussed and the algorithm for selection and convertion of GAO expressions into the set of WACFG productions is provided. The proposed algorithm turned out to be feasible in the context of parsing and translating Ukrainian Spoken and Ukrainian Sign language. The developed approach for mixed semantical and syntactical sentence parsing was tested on the database of sentences from Ukrainian fairy tail by Ivan Franko “Fox Mykyta” where 92 % of sentences were correctly parsed.