Knowledge Retrieval and Relation Mining from Tolkien’s History of Middle Earth

Abstract

In this paper, we focus on the problem of automatic analysis of literature via its characters’ interaction networks. We demonstrate the ability of the well-known knowledge retrieval algorithms on the text series of books by John Ronald Reuel Tolkien about Middle-earth. We study the identified relationships of characters, visualizing the results using the network science approaches via introducing the general pipeline for named entities recognition, relation retrieval, and analysis of their clustering via racial division by using the Adjusted Rand Index.