Choisir la meilleure configuration d'un système de recherche d'information (notice n° 370412)
[ vue normale ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 02521cam a2200289 4500500 |
| 005 - DATE AND TIME OF LATEST TRANSACTION | |
| control field | 20250118081726.0 |
| 041 ## - LANGUAGE CODE | |
| Language code of text/sound track or separate title | fre |
| 042 ## - AUTHENTICATION CODE | |
| Authentication code | dc |
| 100 10 - MAIN ENTRY--PERSONAL NAME | |
| Personal name | Bigot, Anthony |
| Relator term | author |
| 245 00 - TITLE STATEMENT | |
| Title | Choisir la meilleure configuration d'un système de recherche d'information |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Date of publication, distribution, etc. | 2014.<br/> |
| 500 ## - GENERAL NOTE | |
| General note | 78 |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | Les campagnes telles que Text REtrieval Conference (TREC) offrent un cadre qui permet d’évaluer des systèmes de recherche d’information (RI). L’évaluation utilise des mesures qui se basent sur une moyenne des résultats obtenus pour un ensemble de besoins en information : les succès et échecs sur chacun des besoins sont masqués. Dans un premier temps, cet article propose une analyse qui prend en compte la diversité des résultats et qui a pour but de sélectionner le système de RI le plus apte à traiter les besoins selon leur difficulté. L’évaluation montre un gain maximal en robustesse de 68 % et de 10 % sur les performances. Dans un second temps, nous montrons qu’il est possible d’apprendre la meilleure configuration de système en s’appuyant sur un sous-ensemble de documents dont la pertinence est connue. |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | Evaluation campaigns such as Text REtrieval Conference (TREC) offer a framework to evaluate information retrieval (IR) systems. Generally, these evaluation results are averaged over information needs so successes and failures on specific needs are hidden. At first, this article proposes an analysis which takes into account the diversity of results and select the IR system best suited to address the information needs according to their difficulty. An experiment is driven in order to select the best systems according to the information need difficulty. Evaluation results show a maximum gain of 68% in reliability and of 10% on performances. In a second step, we show that it is possible to learn the best system configuration based on a subset of documents which relevance is known. |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | évaluation en RI |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | classifications |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | difficulté des besoins |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | rang des systèmes |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | clustering |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | information need difficulty |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | system ranks |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | IR evaluation |
| 700 10 - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Déjean, Sébastien |
| Relator term | author |
| 700 10 - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Mothe, Josiane |
| Relator term | author |
| 786 0# - DATA SOURCE ENTRY | |
| Note | Document numérique | 17 | 2 | 2014-10-01 | p. 125-147 | 1279-5127 |
| 856 41 - ELECTRONIC LOCATION AND ACCESS | |
| Uniform Resource Identifier | <a href="https://shs.cairn.info/revue-document-numerique-2014-2-page-125?lang=fr&redirect-ssocas=7080">https://shs.cairn.info/revue-document-numerique-2014-2-page-125?lang=fr&redirect-ssocas=7080</a> |
Pas d'exemplaire disponible.
Réseaux sociaux