ViS-\'A-ViS : Detecting Similar Patterns in Annotated Literary Text

TL;DR

ViS-Á-ViS is a web-based tool that helps literary scholars identify repetitive and similar patterns in annotated texts using visualization and dynamic time warping algorithms, demonstrated on Hebrew poetry.

Contribution

The paper introduces a novel system combining visualization and DTW for pattern detection in annotated literary texts, aiding scholarly analysis.

Findings

Effective detection of literary patterns in annotated data

Successful application to Hebrew poetry corpus

Preliminary results show system's usefulness

Abstract

We present a web-based system called ViS-\'A-ViS aiming to assist literary scholars in detecting repetitive patterns in an annotated textual corpus. Pattern detection is made possible using distant reading visualizations that highlight potentially interesting patterns. In addition, the system uses time-series alignment algorithms, and in particular, dynamic time warping (DTW), to detect patterns automatically. We present a case-study where an ancient Hebrew poetry corpus was manually annotated with figurative language devices as metaphors and similes and then loaded into the system. Preliminary results confirm the effectiveness of the system in analyzing the annotated data and in detecting literary patterns and similarities.