Skip to Main content Skip to Navigation
Conference papers

GO-SPADE: Mining Sequential Patterns over Datasets with Consecutive Repetitions

Abstract : Databases of sequences can contain consecutive repetitions of items. This is the case in particular when some items represent discretized quantitative values. We show that on such databases, a typical algorithm like the SPADE algorithm tends to loose its efficiency. SPADE is based on the used of lists containing the localization of the occurrences of a pattern in the sequences and these lists are not appropriated in the case of data with repetitions. We introduce the concept of generalized occurrences and the corresponding primitive operators to manipulate them. We present an algorithm called GO-SPADE that extends SPADE to incorporate generalized occurrences. Finally we present experiments showing that GO-SPADE can handle sequences containing consecutive repetitions at nearly no extra cost.
Complete list of metadatas

https://hal.archives-ouvertes.fr/hal-01588170
Contributor : Équipe Gestionnaire Des Publications Si Liris <>
Submitted on : Friday, September 15, 2017 - 11:23:33 AM
Last modification on : Wednesday, November 20, 2019 - 2:49:21 AM

Links full text

Identifiers

Citation

Marion Leleu, Christophe Rigotti, Jean-François Boulicaut, G Euvrard. GO-SPADE: Mining Sequential Patterns over Datasets with Consecutive Repetitions. International Workshop on Machine Learning and Data Mining in Pattern Recognition, MLDM'03, Jul 2003, Leipzig, Germany. pp.293-306, ⟨10.1007/3-540-45065-3_26⟩. ⟨hal-01588170⟩

Share

Metrics

Record views

100