Mining minimal non-redundant association rules using frequent itemsets lattice
Bay Vo, Bac Hoai Le · International Journal of Intelligent Systems Technologies and Applications · 2011
In this paper, we present an application of frequent itemsets lattice (FIL) to mine minimal non-redundant association rules (MNARs) that reduces a lot of time for mining rules. Our method includes two phases: (1) building FIL and (2) mining MNARs from lattice. we extend the structure of FIL by adding one field to consider whether a lattice node is a minimal generator (mG) or not, and another field to consider whether a lattice node is a frequent closed itemset or not. MNARs are only mined from a minimal generator X to frequent closed itemset Y such that X⊂Y. The experiments show that the mining time from FIL is more effective than that from frequent closed itemsets.