Skip to Main content Skip to Navigation
Conference papers

Détecter le besoin d'information dans des requêtes d'usagers d'agents virtuels : sélection de données pertinentes

Abstract : Selecting relevant data for information need detection in virtual agent user queries Customer relationship platforms offer a variety of communication channels, which include virtual agents (VA). Efficient routing of messages must take into account the user's information need. For a task of classification by information need, it may be useful first to select from these often noisy messages those units of text which relevantly represent the information need. We describe a corpus of VA user queries, and experiment on it with two methods for selecting relevant segments: one based on term extraction, the other on filtering. The results are encouraging but there is room for improvement, and extrinsic evaluation remains to be done.
Complete list of metadatas

Cited literature [16 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01347087
Contributor : Fabienne Moreau <>
Submitted on : Wednesday, July 20, 2016 - 12:11:14 PM
Last modification on : Monday, March 2, 2020 - 9:44:06 AM

File

taln2016_article_court_efraim_...
Files produced by the author(s)

Identifiers

  • HAL Id : hal-01347087, version 1

Collections

Citation

Octavia Efraim, Fabienne Moreau. Détecter le besoin d'information dans des requêtes d'usagers d'agents virtuels : sélection de données pertinentes. 23ème Conférence sur le Traitement Automatique des Langues Naturelles, Jul 2016, INALCO, PARIS, France. ⟨hal-01347087⟩

Share

Metrics

Record views

81

Files downloads

158