Methods of Automatized Semantic Analysis of Natural-Language Information

Apply

Methods of Automatized Semantic Analysis of Natural-Language Information

Journal Title: Філологічний дискурс - Year 2018, Vol 0, Issue 7

Abstract

In the article the features of automated semantic analysis of the text have been considered; problems of creation of automated linguistic-software tools, which are suitable for application in extraction systems of semantics from the text have been investigated. The author has systematized and carried out a comparative analysis of the results of scientific research in the field of developing models and methods of semantic analysis of natural-language information. Two basic approaches in the field of computer processing of natural-language texts have been determined: lingua-analytical and statistical. The author has argued that the most promising and effective of them are, respectively, explicit methods of semantic analysis of text information (algorithms of ontological semantic analysis) and methods of latent semantic analysis. Possible ways of improvement of existing computer means for diagnosing the relevance of natural-language information have been outlined; it has been substantiated that the most popular methods of processing natural-language information for the purpose of extraction and representation of semantics should be systems based on the efficient combination of linguistic analysis technologies (graphical, morphological, syntactic, semantic), in particular using ontologies, and the method of latent semantic analysis. It has been proved that the integration of technologies of explicit semantic analysis, latent semantic analysis, methods of the theory of fuzzy logic, artificial intelligence, etc. is a promising way of solving the problem of automated semantic analysis of natural-language information.

Authors and Affiliations

Оксана Комарницька

Keywords

EP ID EP648418
DOI -
Views 132
Downloads 0