一种关联规则挖掘的高效更新算法

摘要

针对关联规则挖掘中的高效更新问题,对增量和负增量问题进行了讨论,提出当最小支持度发生变化时可归结为数据库发生变化的情形进行讨论。采用十字链表来分别存储原数据库DB和变化数据库db中,各频繁项集及其支持度s1和s2,通过对s1,s2及最小支持度s0的比较分析,判断项集是否为频繁项集,减少了扫描数据库的次数,提高了更新后的挖掘效率。并通过仿真实验,证明了算法的高效性。%Both incremental and negative incremental field are discussed aiming at the problem of mining association rules effectively.It is suggested that the problem of change of minimum support can be transformed to the problem of updating database.The crossing lists are used to storage all frequent items and their supports in both the original and the changed database.After analyzing the relation between s1,s2 and s0,whether an item is frequent or not can be figured out.By this method,the I／O time is largely reduced,and the efficiency of mining association rules has been improved.Finally,the emulation experiment verifies its effectiveness.