Variations axiomatiques pour la recherche d'information personnalisée

Abstract : This paper focuses on difficulty of finding relevant information in a social network. We start from the premise that user profiles must be filtered to have a positive impact on the retrieval. The Profile Query Expansion Constraint is then defined: it defines the expected behavior of the terms that are used to personalize a user query. We define, then, one integration of the constraint in a probabilistic matching framework, before studying when the query may be used to help focusing the social profile of a user asking a query. Experiments are processed on the Bibsonomy corpus. Our findings show that even simple profile adaptations are effective for social personalized information retrieval.
Document type :
Conference papers
Complete list of metadatas

Cited literature [7 references]  Display  Hide  Download

https://hal-ujm.archives-ouvertes.fr/ujm-01615861
Contributor : Mathias Géry <>
Submitted on : Thursday, October 12, 2017 - 5:23:55 PM
Last modification on : Tuesday, February 12, 2019 - 1:31:00 AM

File

2017_03_21_CORIA2017_paper_12_...
Files produced by the author(s)

Identifiers

  • HAL Id : ujm-01615861, version 1

Citation

Philippe Mulhem, Nawal Ould Amer, Mathias Géry. Variations axiomatiques pour la recherche d'information personnalisée. COnférence en Recherche d'Informations et Applications, Mar 2017, Marseille, France. ⟨ujm-01615861⟩

Share

Metrics

Record views

263

Files downloads

32