Constrained multidimensional patterns differ from the well-known frequent patterns from a conceptual and logical points of view because they are provided with a common structure and support various types of constraints. Classical data mining techniques are based on the power set lattice of binary attributes and, even extended, are not suitable when addressing the discovery of constrained multidimensional patterns. In this paper we propose a foundation for various multidimensional data mining problems by introducing a new algebraic structure called cube lattice which characterizes the search space to be explored. We take into consideration monotone and/or antimonotone constraints enforced when mining multidimensional patterns. In addition, we propose condensed representations of the constrained cube lattice which is a convex space. Finally, we place emphasis on advantages of the cube lattice when compared to the power set lattice of binary attributes used for multidimensional data mining.
[1]
Heikki Mannila,et al.
Levelwise Search and Borders of Theories in Knowledge Discovery
,
1997,
Data Mining and Knowledge Discovery.
[2]
Lotfi Lakhal,et al.
Treillis Relationnel : Une Structure Algébrique pour le Data Mining Multi-Dimensionnel
,
2002,
BDA.
[3]
Gregory Piatetsky-Shapiro,et al.
Advances in Knowledge Discovery and Data Mining
,
2004,
Lecture Notes in Computer Science.
[4]
Heikki Mannila,et al.
Fast Discovery of Association Rules
,
1996,
Advances in Knowledge Discovery and Data Mining.
[5]
Ramakrishnan Srikant,et al.
Mining quantitative association rules in large relational tables
,
1996,
SIGMOD '96.
[6]
Thomas G. Dietterich.
What is machine learning?
,
2020,
Archives of Disease in Childhood.
[7]
Van de M. L. J. Vel.
Theory of convex structures
,
1993
.