Identifying hidden communities of interest with topic-based networks: A Case study of the community of philosophers of science (1930-2017).

Malaterre, C et Lareau, F (2023). « Identifying hidden communities of interest with topic-based networks: A Case study of the community of philosophers of science (1930-2017). », dans Proceedings of the 56th Hawaii International Conference on System Sciences (HICSS) (56th Hawaii International Conference on System Sciences (HICSS, Hawaii, 2023-01-03) Hawaii, pp. 2473-2483.

Fichier(s) associé(s) à ce document :
[img]
Prévisualisation
PDF
Télécharger (6MB)

Résumé

Scientific networks are often investigated by means of citation analyses. Yet, interpretation of such networks in terms of semantic (and often disciplinary) content heavily depends on supplementary knowledge, notably about author research specialties. Similar situations arise more generally in many types of social networks whose semantic interpretation relies on supplementary information. Here, author community net-works are inferred from a topic model which provides direct insights into the semantic specificity of the identified “hidden communities of interest” (HCoI). Using a philosophy of science corpus of full-text articles (N=16,917), we investigate its underlying communities by measuring topic profile correlations be-tween authors. A diachronic perspective is built by modeling the research networks over different time-periods and mapping genealogical relationships be-tween communities. The results show a marked in-crease in philosophy of science communities over time and trace the progressive appearance of the specialization areas that structure the field today.

Type: Communication, article de congrès ou colloque
Mots-clés ou Sujets: Network Analysis of Digital and Social Media, hidden communities of interest, philosophy of science, social networks, text-mining, topic model
Unité d'appartenance: Faculté des sciences humaines > Département de philosophie
Déposé par: Christophe Malaterre
Date de dépôt: 17 avr. 2026 08:13
Dernière modification: 17 avr. 2026 08:13
Adresse URL : https://archipel.uqam.ca/secure/id/eprint/19706

Statistiques

Voir les statistiques sur cinq ans...