Co-occurrence matrices of time series applied to literary works

TL;DR

This paper introduces a method using co-occurrence matrices over time to analyze the presence and relevance of characters in literary works, offering a new way to incorporate timelines into social network analysis.

Contribution

It presents a novel approach that applies co-occurrence matrices to time series data in literary analysis, integrating timelines with network visualization.

Findings

Co-occurrence matrices reveal character interactions over time.

The method captures the dynamics of character presence.

Matrices resemble recurrence plots for time series analysis.

Abstract

Recently, it has been proposed to analyse the literary works, plays or novels, using graphs to display the social network of their interacting characters. In this approach, the timeline of the literary work is lost, because the storyline is projected on a planar graph. However, timelines can be used to build some time series and analyse the work by means of vectors and matrices. These series can be used to describe the presence and relevance, not only of words in the text, but also of persons and places portrayed in the drama or novel. In this framework, we discuss here an approach with co-occurrence matrices plotted over time, concerning the presence of characters in the pages of a novel. These matrices are similar to those appearing in recurrence plots.