Digital Library

Detrended fluctuation analysis of sentiment patterns in literary texts

• Lukasz Gaza
• Stanislaw Drozdz
• Pawel Oswiecimka
• Marek Stanuszek

(2023). ECMS 2023, 37th Proceedings
Edited by: Enrico Vicario, Romeo Bandinelli, Virginia Fani, Michele Mastroianni, European Council for Modelling and Simulation.
DOI: http://doi.org/10.7148/2023
ISSN: 2522-2422 (ONLINE)
ISSN: 2522-2414 (PRINT)
ISSN: 2522-2430 (CD-ROM)
ISBN: 978-3-937436-80-7
ISBN: 978-3-937436-79-1 (CD) Communications of the ECMS Volume 37, Issue 1, June 2023, Florence, Italy June 20th – June 23rd, 2023

https://doi.org/10.7148/2023-0549

Lukasz gaza, Stanislaw drozdz, Pawel oswiecimka, Marek stanuszek (2023). Detrended fluctuation analysis of sentiment patterns in literary texts, ECMS 2023, Proceedings Edited by: Enrico Vicario, Romeo Bandinelli, Virginia Fani, Michele Mastroianni, European Council for Modelling and Simulation. doi:10.7148/2023-0549

Temporal correlations of the sentiment content in consecutive sentences is studied based on a large corpus of world-famous literary text in four major European languages (English, French, German and Spanish). For quantifying the related characteristics in terms of the Hurst exponents the Detrended Fluctuation Analysis (DFA) is employed. The results obtained provide a clear indication for the existence of persistent correlations as revealed by the Hurst exponents significantly larger than 1/2 in all the four languages studied. An extremely interesting and worth further more systematic study is the identified fact that in many texts the DFA indicates two different regimes of scaling and thus two different Hurst exponents. In those cases - quite universally - the cross-over occurs at the scale corresponding to about 200 sentences and the Hurst exponents at the scales above this value are even larger which indicates the presence of stronger long-range temporal correlations than those at the shorter scales.