Clustering incrémental et méthodes de détection de nouveauté : application à l'analyse intelligente d'informations évoluant au cours du temps

Abstract : Learning algorithms proved their ability to deal with large amount of data. Most of the statistical approaches use defined size learning sets and produce static models. However in specific situations: active or incremental learning, the learning task starts with only very few data. In that case, looking for algorithms able to produce models with only few examples becomes necessary. The literature's classifiers are generally evaluated with criteria such as: accuracy, ability to order data (ranking)... But this classifiers' taxonomy can really change if the focus is on the ability to learn with just few examples. To our knowledge, just few studies were performed on this problem. This study aims to study a larger panel of both algorithms (9 different kinds) and data sets (17 UCI bases).
Liste complète des métadonnées

Cited literature [73 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-00962376
Contributor : Patricia Gautier <>
Submitted on : Wednesday, March 26, 2014 - 2:30:51 PM
Last modification on : Tuesday, December 18, 2018 - 4:38:01 PM
Document(s) archivé(s) le : Thursday, June 26, 2014 - 10:46:03 AM

File

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

Identifiers

  • HAL Id : hal-00962376, version 1

Collections

Citation

Pascal Cuxac, Jean-Charles Lamirel. Clustering incrémental et méthodes de détection de nouveauté : application à l'analyse intelligente d'informations évoluant au cours du temps. La recherche d'information en contexte :Outils et usages applicatifs, L. Grivel Ed, pp.00, 2011. ⟨hal-00962376⟩

Share

Metrics

Record views

448

Files downloads

2781