AMPHAWAN Komate1,2,3, LENCA Philippe1,2, SURARERKS Athasit3
Chapitre dans un livre
New Frontiers in Applied Data Mining - PAKDD 2011 International Workshops, Shenzhen, China, May 24-27, 2011, Revised Selected Papers, Springer-Verlag Berlin Heidelberg, 2012, (Lecture Notes in Computer Science, 7104), pp. 124-135
Association rule discovery based on support-confidence framework is an important task in data mining. However, the occurrence frequency (support) of a pattern (itemset) may not be a sufficient criterion for discovering interesting patterns. Temporal regularity, which can be a trace of behavior, with frequency behavior can be revealed as an important key in several applications. A pattern can be regarded as a regular pattern if it occurs regularly in a user-given period. In this paper, we consider the problem of mining top-k regular-frequent itemsets from transactional databases without support threshold. A new concise representation, called compressed transaction-ids set (compressed tidset), and a single pass algorithm, called TR-CT (Top-k Regular frequent itemset mining based on Compressed Tidsets), are proposed to maintain occurrence information of patterns and discover k regular itemsets with highest supports, respectively. Experimental results show that the use of the compressed tidset representation achieves highly efficiency in terms of execution time and memory consumption, especially on dense datasets.
1 : LUSSI - Dépt. Logique des Usages, Sciences Sociales et de l'Information (Institut Mines-Télécom-Télécom Bretagne-UEB)
2 : Lab-STICC - Laboratoire en sciences et technologies de l'information, de la communication et de la connaissance (UMR CNRS 6285 - Télécom Bretagne - Université de Bretagne Occidentale - Université de Bretagne Sud)
3 : ELITE - Engineering Laboratory in Theoretical Enumerable System (University of Chulalongkorn)
Frequent itemsets, Regular itemsets, Top-k itemsets