An Incremental Updating Algorithm for Online Mining Association Rules

Abstract

Aimed at the limitation of the current FUP algorithm, which readily led to the low efficiency of frequent itemset of the updated database, a novel association rules mining algorithm named QAIS which is different from the classical dual phrase was proposed. On the basis of QAIS, an improved association rules mining algorithm named AIU was put forward further. AIU can efficiently maintain association rules when the transaction database was updated. At the same time, the minimum support and confidence were not changed. Experiments showed the proposed algorithm was appropriate to online association rules mining.