Identifying Latin authors through maximum-likelihood Dirichlet inference: A contribution to model-based stylometry (original) (raw)

Shagi / Steps, 2021

Abstract

We present a new algorithm for stylometric identification of authors of Latin prose texts based on Burrows' Delta and the Dirichlet distribution. In order to demonstrate the effectiveness of the method, we present a case study of fragments of texts by 36 authors and show that our method performs on par with Random Forest, a powerful general-purpose classification algorithm. The advantages of our method are that it is very simple to implement, much quicker to train and do inference with, and that it is more interpretable since it is a model-based algorithm: precision of the fitted Dirichlet distributions directly corresponds to the stylistic homogeneity of the texts by different authors.

Mikhail Shumilin hasn't uploaded this paper.

Let Mikhail know you want this paper to be uploaded.

Ask for this paper to be uploaded.