Language attribution of an unmarked text corpus

Abstract

Unmarked text corps will increasingly appear with the growth of information on the web. Automated analysis of Big Data in search engines, scientific and commercial applications requires detailed information about the object under study. In the case of text bodies, information on the language of the documents is extremely important. Working with the scanned texts the situation is even more complicated. In this paper, the idea of using the fractal-inspired irregularity to attribute the language of the text is being further developed. A methodology for the attribution is proposed and an experiment based on 10 European languages is conducted. The proposed approach has shown its effectiveness and promise. A selection of approximately 4000 characters (1 page of text) allows you to uniquely attribute the language of the text. © 2020, World Scientific and Engineering Academy and Society. All rights reserved.