Indexation de co-occurrences dans des corpus de documents structurés et production de cartes sémantiques interactives (notice n° 402140)
[ vue normale ]
| 000 -LEADER | |
|---|---|
| fixed length control field | 03640cam a2200313 4500500 |
| 005 - DATE AND TIME OF LATEST TRANSACTION | |
| control field | 20250119075044.0 |
| 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 | Pompidor, Pierre |
| Relator term | author |
| 245 00 - TITLE STATEMENT | |
| Title | Indexation de co-occurrences dans des corpus de documents structurés et production de cartes sémantiques interactives |
| 260 ## - PUBLICATION, DISTRIBUTION, ETC. | |
| Date of publication, distribution, etc. | 2010.<br/> |
| 500 ## - GENERAL NOTE | |
| General note | 68 |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | RésuméConfrontés à la problématique de l’indexation de très grands corpus documentaires d’entreprises, nous avons mis au point une méthode simple mais efficace (en temps de calcul et de volumétrie), permettant de filtrer par document les co-occurrences les plus représentatives de ceux-ci. Le choix d’un contexte de co-occurrences a deux raisons. D’une part les requêtes portant sur des corpus spécialisés et composés par des experts, s’appuient sur peu de termes précisément choisis dont l’indexation des associations permet la construction de cartes sémantiques de navigation dans les concepts du corpus. Pour cela nous prenons en compte la structure des documents en validant les contenus des paragraphes par ceux de leurs titres. Notre méthode s’appuie sur des mesures tf.idf successives effectuées dans le contexte d’un document et non d’un corpus, sur les contenus des paragraphes auxquels sont intégrés progressivement la hiérarchie des titres les introduisant. D’autre part, nous exploitons simultanément une ontologie de contrôle et les requêtes des utilisateurs comportant les termes précédemment discriminés pour valider par le théorème de Bayes, les associations sémantiques ainsi déterminées, qui finalement permettent la production de cartes sémantiques. |
| 520 ## - SUMMARY, ETC. | |
| Summary, etc. | This paper addresses the problem of indexing very large enterprise corpuses. We have designed a simple yet efficient (in terms of computation time and the size of the generated results) method allowing to filter, on a per-document basis, the most representative co-occurrences of the documents. The reason for using co-occurrences is twofold. First, queries composed by experts on specialized corpuses rely statistically on few chosen terms, for which we index the associations. Second, such co-occurrences facilitate the construction of semantic maps used to navigate the concepts of the corpus. Our main approach is to take into account the structure of the documents by validating the content of the paragraphs by their titles. Our method starts with successive tf.idf measures of paragraph contents taken in the context of a document, to which we progressively integrate the hierarchy of their introducing titles. We then simultaneously exploit a control ontology and the user queries containing the terms that we discriminated in the first step in order to validate, using Bayes’ theorem, the semantic associations contained in a paragraph given the terms of its title.DOI:10.3166/DN.12.1.53-79 © 2009 Lavoisier, Paris |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | contexte de co-occurrences |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | indexation incrémentale de très grands corpus |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | théorème de Bayes |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | exploitation de la structure des documents |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | cartes sémantiques |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | Bayes' theorem |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | co-occurrences' context |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | semantic maps |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | very large and fast corpus indexing |
| 690 ## - LOCAL SUBJECT ADDED ENTRY--TOPICAL TERM (OCLC, RLIN) | |
| Topical term or geographic name as entry element | structured document retrieval |
| 700 10 - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Carbonneill, Boris |
| Relator term | author |
| 700 10 - ADDED ENTRY--PERSONAL NAME | |
| Personal name | Sala, Michel |
| Relator term | author |
| 786 0# - DATA SOURCE ENTRY | |
| Note | Document numérique | 12 | 1 | 2010-01-18 | p. 53-79 | 1279-5127 |
| 856 41 - ELECTRONIC LOCATION AND ACCESS | |
| Uniform Resource Identifier | <a href="https://shs.cairn.info/revue-document-numerique-2009-1-page-53?lang=fr&redirect-ssocas=7080">https://shs.cairn.info/revue-document-numerique-2009-1-page-53?lang=fr&redirect-ssocas=7080</a> |
Pas d'exemplaire disponible.
Réseaux sociaux