A Hierarchy-Based Method for Synthesizing Frequent Itemsets Extracted from Temporal Windows

Yoann Pitarch 1 Anne Laurent 1 Pascal Poncelet 1
1 TATOO - Fouille de données environnementales
LIRMM - Laboratoire d'Informatique de Robotique et de Microélectronique de Montpellier
Abstract : With the rapid development of information technology, many applications have to deal with potentially infinite data streams. In such a dynamic context, storing the whole data stream history is unfeasible and providing a high-quality summary is required for decision makers. A practical and consistent summarization method is the extraction of the frequent itemsets over temporal windows. Nevertheless, this method suffers from a critical drawback: results pile up quickly making the analysis either uncomfortable or impossible for users. In this paper, we propose to unify these results thanks to a synthesis method for multidimensional frequent itemsets based on a graph structure and taking advantage of the data hierarchies. We overcome a major drawback of the Tilted Time Window (TTW) standard framework by taking into account the data distribution. Experiments conducted on both synthetic and real datasets show that our approach can be applied to data streams.
Type de document :
Communication dans un congrès
SoCPaR'09: The International Conference of Soft Computing and Pattern Recognition, Dec 2009, pp.001-006, 2009, 〈http://www.mirlabs.org/socpar/index.html〉
Liste complète des métadonnées

https://hal-lirmm.ccsd.cnrs.fr/lirmm-00426492
Contributeur : Yoann Pitarch <>
Soumis le : lundi 26 octobre 2009 - 14:56:39
Dernière modification le : jeudi 11 janvier 2018 - 06:26:17

Identifiants

  • HAL Id : lirmm-00426492, version 1

Collections

Citation

Yoann Pitarch, Anne Laurent, Pascal Poncelet. A Hierarchy-Based Method for Synthesizing Frequent Itemsets Extracted from Temporal Windows. SoCPaR'09: The International Conference of Soft Computing and Pattern Recognition, Dec 2009, pp.001-006, 2009, 〈http://www.mirlabs.org/socpar/index.html〉. 〈lirmm-00426492〉

Partager

Métriques

Consultations de la notice

28