Les topic models au service de l’histoire d’un genre vidéoludique. Vers une représentation non périodique de l’évolution du contenu textuel des jeux de rôle sur ordinateur entre 1992 et 2017

Translated title of the contribution: Using Topic Models to Study the History of a Video Game Genre: Towards a Non-Periodic Representation of Changes in the Textual Content of Computer Role-Playing Games between 1992 and 2017

Fanny Barnabé, Nicolas Bourgeois

Research output: Contribution to journalArticlepeer-review

Abstract

Using Topic Models to Study the History of a Video Game Genre: Towards a Non-Periodic Representation of Changes in the Textual Content of Computer Role-Playing Games between 1992 and 2017. This paper shows how researchers can combine and exploit mathematical tools still rarely used by historians (topic models, hierarchical classification, self-organizing maps) to study a dated but heterogeneous historical corpus and to characterize its evolution over time. The analysis focuses on changes in the vocabulary used by a set of 21 popular western role-playing video games published between 1992 and 2017 and comprising a total of 17.5 million words. Through this example, we aim to shed new light on the canonical history of the genre and to propose an alternative model to periodization for uncovering the trade-offs and influences underlying the development of each particular game must negotiate with.

Translated title of the contributionUsing Topic Models to Study the History of a Video Game Genre: Towards a Non-Periodic Representation of Changes in the Textual Content of Computer Role-Playing Games between 1992 and 2017
Original languageFrench
Pages (from-to)199-233
Number of pages35
JournalHistoire et Mesure
Volume36
Issue number2
DOIs
Publication statusPublished - 1 Jul 2022
Externally publishedYes

Fingerprint

Dive into the research topics of 'Using Topic Models to Study the History of a Video Game Genre: Towards a Non-Periodic Representation of Changes in the Textual Content of Computer Role-Playing Games between 1992 and 2017'. Together they form a unique fingerprint.

Cite this