Computational Literary Analysis of Narrative Structures and Character Development Using Natural Language Processing and Machine Learning
Computer and Decision Making: An International Journal · 04.11.2026 · 3 (6S) · 2869-2878
DOI: 10.1922/comdem.v3i6s.19040 ↗Kurz gefasst
Fachartikel in der Zeitschrift Computer and Decision Making: An International Journal, online erschienen am 04.11.2026. Verfasst von Dr. Apoorva Sinha und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Literary narratives are now beginning to be analyzed using measurable linguistic, emotional, structural, and relational patterns, which is made possible by computational methods. This research delves into the structure of narratives and character development analysis, employing natural language processing, network analysis, and machine learning techniques. There were 108,093 records across 37 plays in a source corpus that were processed for computational analysis. Linguistic features, lexical diversity, sentiment polarity, thematic composition, dialogue participation, and character-network measures were analysed across successive narrative stages. Character development was captured by changes in linguistic, emotional, thematic and interaction features over time. The two general profiles of character development that emerged from the unsupervised clustering analysis were distinguished primarily by the extent to which the characters engage in dialogue, are central to the network, feel emotions that vary across their utterances, and experience shifts in their textual themes. Analysis of the narratives showed that the sentiment of the narratives tended to fade in the later acts, that there were emotional "arcs" in the narratives for each genre, that the thematic focus shifted, and that the density of the character networks grew. Additionally, supervised classification was performed to show that combined textual and structural features could be used to classify literary genres. The random forest model performed best, with an accuracy of 0.7622 and a macro-F1 score of 0.7585, which is higher than the accuracy of the support vector machine model and the logistic regression model. The results show the possibility of depicting narrative development and character evolution as a multidimensional process of language, emotions, themes and relationships, as well as their interconnections. The framework of computation outlines a systematic process for discovering repeated patterns of narrative, character development, and genres in literary texts.
Zitieren
APA
Sinha, D. A., Riba, D. K., Bandameedi, S., Balan, D. N. S., & Kishor, P. (2026). Computational Literary Analysis of Narrative Structures and Character Development Using Natural Language Processing and Machine Learning. Computer and Decision Making: An International Journal, 3(6S), 2869-2878. https://doi.org/10.1922/comdem.v3i6s.19040
Vancouver
Sinha DA, Riba DK, Bandameedi S, Balan DNS, Kishor P. Computational Literary Analysis of Narrative Structures and Character Development Using Natural Language Processing and Machine Learning. Computer and Decision Making: An International Journal. 2026;3(6S):2869-2878. doi:10.1922/comdem.v3i6s.19040
BibTeX
@article{Sinha2026,
author = {Sinha, Dr. Apoorva and Riba, Dr Karyir and Bandameedi, Srinivas and Balan, Dr Neema Sajju and Kishor, Prof.Nand},
title = {Computational Literary Analysis of Narrative Structures and Character Development Using Natural Language Processing and Machine Learning},
journal = {Computer and Decision Making: An International Journal},
year = {2026},
volume = {3},
number = {6S},
pages = {2869--2878},
doi = {10.1922/comdem.v3i6s.19040}
}
EUVIMED