Mining Concise Representations of Frequent Multidimensional Patterns

摘要

In this paper, we consider an instance of the constraint pattern problem: the frequent multidimensional patterns. We propose the first concise representation of frequent multidimensional patterns which on one hand is not based on the powerset lattice framework and on the other hand differs from the representations by closed patterns and non derivable patterns. From such a representation, we show that any frequent multidimensional pattern along with its conjunction, disjunction and negation frequencies can be obtained using inclusion-exclusion identities.