Searching Data Cube for Submerging and Emerging Cuboids

Viet Phan-Luong (1, 2, 3)

1 AMU - Aix Marseille Université
2 LIF - Laboratoire d'informatique Fondamentale de Marseille
3 DANA - Data Mining at scale

—Many existing approaches to data cube computation search for the group-by partitions on fact table with support greater than some threshold, that is, those can be obtained from the SQL group-by queries with the clause HAVING COUNT(*) >= supp, where supp is a support threshold. Those partitions constitute what is called the iceberg data cube. The present work proposes an efficient method to compute the similar partitions, but in contrast to those previous approaches, in the clause HAVING, COUNT(*) can be >= supp or < supp. Generally, the method can search for the partitions that can be in the emerging part or in the submerging part of data cube, and on any other aggregate functions, not only the function COUNT.

Mots clés: Iceberg data cube; Data mining; Data warehouse; Infrequent itemset

Domaines: Base de données [cs.DB]

Viet Phan-Luong. Searching Data Cube for Submerging and Emerging Cuboids. 2017 IEEE 31st International Conference on Advanced Information Networking and Applications (AINA), Mar 2017, Taipei, France. ⟨10.1109/AINA.2017.77⟩. ⟨hal-01796021⟩