Mining cyclic association rules from multidimensional knowledge

The cyclicity search in temporal databases aims at discovering regularly repeated rules over time. Despite the increasing number of the proposed works handling mining cyclic association rules, few works paid attention to generating rules like ("A company producing a given product will sell it with respect to such sales amount. This fact is repeated each month"). In this paper, we propose a new definition of cyclic association rules extracted from several dimensions. Thus, we introduce a new algorithm RACYM for generating such rules. Carried out experiments, conducted on a real data warehouse, show the usefulness and the performance of our proposal.