N-gram Web Service and Stylometric Analysis of Korean Historical Documents

Abstract

The information revolution enables us to access almost everything in the world simply online. Thanks to the internet, people now have easy access to old books, which previously had only been available to a few selected experts in specific fields. In this paper, we examine five historical records ('The Annals of the Joseon Dynasty', 'The Daily Records of the Royal Secretariat of the Joseon Dynasty', 'Records of the Border Defense Council', 'Goryeo History', and 'Essentials of Goryeo History') of Korea before the Japanese colonial era from the perspective of stylometry. First, we construct a web service that shows how often a particular Chinese character appears in the old books across time. We compare the frequency distributions of the Chinese characters used in these five different records. We also study the stylistic similarity by using the yearly frequency distribution of content-free characters. This work may provide macroscopic and systematic insights into large-scale analyses of historical records.