基于多事务的关联规则增量更新

在多次挖掘结果上进行增量更新可以显著地降低对原始交易数据库的访问,但一般都认为这是一种以空间换时间的方法。文章指出,只需要增加极少量的存储空间就可以存储同一交易数据库多个最小支持度下的所有频繁项目集,并给出相应的表示模型。在此基础上系统地研究了在已有的多次挖掘结果上的增量更新问题。理论分析和实验结果都表明,该文的方法充分利用已有的多次挖掘结果,降低了算法的复杂度。

Incrementally Updating Association Rules Based on Multiple Tracactions

Incrementally updating association rules based on two or more classes of frequent item sets may reduce the cost of scanning the original database remarkably.However,it is considered as a method of saving time with more storage space.It is put forward in this paper that all frequent item sets of the original database in a serial of minimal supports can be store in a table with a little additional storage.A representation model is given.Based on this model,this paper systematically discusses the problem of incremental updating on discovered association rules of several minimal supports.Tests show that the approach takes full advantage of the previous results and reduces the complexity of incremental updating algorithms.