基于项位置索引的闭合连续序列模式挖掘算法

Closed continuous sequence pattern mining algorithm based onitem position index

为解决现有的闭合序列挖掘算法挖掘大量低效和冗余的序列模式导致的挖掘效率低下,利用频繁项的位置索引,通过索引位置向后扩展频繁连续局部候选项来增长候选序列,采用向前与向后扫描两种策略检查当前候选序列的闭合性,提出了一种具有连续约束的基于频繁项位置索引的闭合连续序列模式挖掘算法LCCS(located closed contiguous sequence data mining)。实验结果表明,LCCS算法通过直接定位频繁项的位置降低数据库的扫描次数,提高了算法挖掘效率,并且得到具有连续约束的闭合序列,有效地解决了产生大量冗余模式的缺陷。

In order to solve the low efficiency problem of mining a large number of inefficient and redundant sequence patterns caused by the existing closed sequence mining algorithms,this paper uses the location index of frequent items to grow candidate sequences by extending the index location backward to frequent continuous local candidates,uses forward and backward scanning strategies to check the closure of current candidate sequences,and proposes a continuous constrained Located Closed Continuous Sequence Data Mining(LCCS)algorithm based on the frequent item position index.The experimental results show that the LCCS algorithm reduces the number of database scans by directly locating the frequent items,improves the efficiency of algorithm mining,and obtains a closed sequence with continuous constraints,which effectively solves the problem of generating a large number of redundant patterns.