Word embedding of religious scripts -

Abstract

Based on the idea that “a word is characterized by the company it keeps”(Firth, 1957) word embedding attempt to derive a set of vectors to represent the corpus vocabulary. Thus having a new embedded space where the geometry between vectors capture the semantic relationship between the words. In this paper, we investigate and demonstrate how the distributed representation resulted by word embedding can be leveraged to quantify and show differences between religious books (Torah, Bible and Qur’an). And explore how the translation of such books affect such automated representations.