Bandit Contextuel pour la Capture de Données Temps Réel sur les Médias Sociaux

Thibault Gisselbrecht 1, 2, * Sylvain Lamprier 2 Patrick Gallinari 2
* Corresponding author
2 MLIA - Machine Learning and Information Access
LIP6 - Laboratoire d'Informatique de Paris 6
Abstract : Social media usually provide streaming data access that enable dynamic capture of the social activity of their users. Leveraging such APIs for collecting data that satisfy a given pre-defined need may constitute a complex task, that implies careful stream selections. On large social media, this represents a very challenging task due to the huge number of potential targets, the intrinsic non-stationarity of user's behavior, and restricted access to the data. We propose an approach that anticipates which profiles are likely to publish relevant contents and dynamically selects a subset of accounts to follow at each iteration using a contextual bandit algorithm. We conduct experiments on Twitter that demonstrate the empirical effectiveness of our approach in real-world settings.
Liste complète des métadonnées

Cited literature [10 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01355408
Contributor : Thibault Gisselbrecht <>
Submitted on : Tuesday, August 23, 2016 - 2:00:25 PM
Last modification on : Thursday, March 21, 2019 - 2:17:50 PM
Document(s) archivé(s) le : Thursday, November 24, 2016 - 12:31:44 PM

File

5.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01355408, version 1

Citation

Thibault Gisselbrecht, Sylvain Lamprier, Patrick Gallinari. Bandit Contextuel pour la Capture de Données Temps Réel sur les Médias Sociaux. Semaine du Document Numérique et de la Recherche d'Information (SDNRI 2016), Mar 2016, Toulouse, France. pp.57-72. ⟨hal-01355408⟩

Share

Metrics

Record views

398

Files downloads

159