Automatic extraction of named entities in the Chinese-Russian corpus of parallel and comparable texts on political topics

Abstract

The aim of the research is to experimentally identify and interpret standard and nested named entities in Chinese and Russian political texts, common and specific to the compared languages, using HanLP and SpaСy libraries. During the study, a Chinese-Russian corpus of parallel and comparable political texts was created. The scientific novelty of the research lies in presenting the results of recognizing various named entities and systematizing the types of errors in the Chinese-Russian corpus of parallel and comparable political texts. The study found that the most frequent named entities in original Chinese and Russian political texts are location names, followed by organization names, with person names being the least frequent. Most high-frequency named entities in original Chinese and translated texts generally correspond to each other, proving that translators often use literal translation when rendering named entities from Chinese into Russian in political texts. Our research systematizes and summarizes information on nested named entities in political texts, identifying and analyzing the following types: [[location]LOCATION], [[location]ORGANIZATION], [[number]ORGANIZATION], [[location]OBJECT], [[location]PROJECT].