Method of Structural Annotations Creation
Y. Platonov, E. Artamonova · 2019 International Multi-Conference on Engineering, Computer and Information Sciences (SIBIRCON) · 2019
The research is dedicated to extracting the semantics from nature language texts applying the big data analysis. In the research, the authors focus on the solution of the particular task of text processing. In the current article authors would like to introduce a new original method of extracting semantics from text in natural language as a completed result of 4-years research. The method was implemented for texts in Russian, but there is a simple way to adapt it for other languages. The method allows working with any text got from any source including html pages. Also the method is suitable for processing linked data or any other type of the data having RDF formatting like triples “object - predicate - subject”. The main idea of the method is to create a syntactic tree - structured annotations - for any sentence in the text, then create logical links between these trees. After that resulting trees are uploaded to DBMS for the future processing and getting results. As DBMS we use the specialized system ORM Polar. We'd chosen it because of its fast data processing and great ability of system scaling.