关联规则的增量更新策略研究

关联规则增量更新算法可以减少对交易数据库的访问。根据最小支持度和交易数据库的不同变化情况，可以将增量更新问题分为若干类。已有的研究只针对某一类具体的增量更新问题，而且没有考虑多次增量更新的情形。本文将增量更新问题归约为三个基本的子问题，从而为各种不同的增量更新问题提供了统一的解决框架。在此基础上，研究了 了多次增量更新事务情况下的增量更新策略问题，通过对增量更新事务进行合理的排列，可以显著地降低对原始交易数据库的访问量。

Research on the Incremental Updating Strategy for Association Rules

Incremental updating algorithms(IUA) for discovered association rules are employed to decrease the cost of scanning the original transaction database.There are several types of IUAs according to different modifications over minimal support and transaction databases.However,the existing research only focuses on a certain type of incremental updating problems(IUPs) and no research has been done to deal with the relationships among all kinds of IUPs.In this paper,all kinds of IUPs are reduced to three basic problems.Based on this,an universal framework for all kinds of IUPs is put forward,and an algorithm which needs not scan the original database is given.Furthermore,this paper presents the concept of incremental updating strategy(IUS) for a series of incremental updating transactions,and the cost of scanning the original database is cut down by rearranging the given incremental updating transactions.