Incremental data mining based on two support thresholds

Abstract

Proposes the concept of pre-large item sets and designs a novel, efficient incremental data mining algorithm based on it. Pre-large item sets are defined using two support thresholds (a lower support threshold and an upper support threshold) to reduce re-scanning of the original databases and to save maintenance costs. The proposed algorithm doesn't need to re-scan the original database until a number of transactions have arrived. If the size of the database is growing larger, then the allowed number of new transactions will be larger too. Therefore, along with the growth of the database, our proposed approach is increasingly efficient. This characteristic is especially useful for real applications.