Дослідження комбінованих методів вилучення структурованих даних для мов з вільним порядком слів

Abstract

This work is devoted to the development of a technique for extracting facts and events from texts, which is based on combining approaches to the use of rules on context-free grammars together with the analysis of syntactic trees of dependency grammars. A combined approach for the problem of fact extraction for languages with free word order is considered. The idea of the technique is to combine an approach to extracting information based on the fusion of results obtained using context-free grammars with the results of pattern selection during text parsing. Based on the proposed approach, a software library has been developed, which includes own templates for extracting subgraphs from dependency trees, considering morphological and syntactic features and the results of extraction using context-free grammars