Extracting temporal patterns from interval-based sequences

Thomas Guyet 1 René Quiniou 2
2 DREAM - Diagnosing, Recommending Actions and Modelling
Inria Rennes – Bretagne Atlantique , IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
Abstract : Most of the sequential patterns extraction methods proposed so far deal with patterns composed of events linked by temporal relationships based on simple precedence between instants. In many real situations, some uantitative information about event duration or inter-event delay is necessary to discriminate phenomena. We propose the algorithm QTIPrefixSpan for extracting temporal patterns composed of events to which temporal intervals describing their position in time and their duration are associated. It extends algorithm PrefixSpan with a multi-dimensional interval clustering step for extracting the representative temporal intervals associated to events in patterns. Experiments on simulated data show that our algorithm is efficient for extracting precise patterns even in noisy contexts and that it improves the performance of a former algorithm which used a clustering method based on the EM algorithm.
Liste complète des métadonnées

https://hal.inria.fr/inria-00618444
Contributor : Thomas Guyet <>
Submitted on : Thursday, September 1, 2011 - 5:54:39 PM
Last modification on : Thursday, November 15, 2018 - 11:57:04 AM

Identifiers

  • HAL Id : inria-00618444, version 1

Citation

Thomas Guyet, René Quiniou. Extracting temporal patterns from interval-based sequences. International Join Conference on Artificial Intelligence (IJCAI), Jul 2011, Barcelone, Spain. ⟨inria-00618444⟩

Share

Metrics

Record views

444