Updating high average-utility itemsets in dynamic databases

Abstract

In this paper, a maintenance algorithm for average-utility mining is proposed to update derived high average-utility itemsets in dynamic databases. It first calculates the count difference of modified itemsets and then partitions them into four parts according to whether they are high upper-bound average-utility itemsets in the original database and whether their count difference is positive or negative. Each part is then processed in its own way. Experimental results show the proposed maintenance algorithm runs faster than the two-phase approach for mining high average-utility itemsets in dynamic databases.