位序互补的空间关联规则挖掘算法

Algorithm of spatial association rules mining based on complementary location order

为有效地提高基于空间事务的挖掘算法效率,提出一种基于位序的互补空间挖掘算法,其适合在海量数据中挖掘任何长度的频繁项;该算法用定序项目集的方法来减少现有算法存在的冗余判断操作和计算,同时也用非频繁项目集的补集来快速生成候选项,删除了现有双向挖掘算法中的空闲操作,可以有效地降低算法的运行时间。实验结果表明此法在空间数据挖掘中比现有算法更快速而有效。

In order to efficiently improve the efficiency of mining algorithms based on spatial transactions,this paper proposes an algorithm of complementary spatial mining based on location order,which is suitable for mining any long frequent itemsets in many data.The algorithm uses the way of locating itemsets＇order to reduce redundant judgment operation and computing existing in presented algorithms,simultaneously uses complement of non-frequent itemsets to fast generate candidate to delete added operation existing in presented double mining algorithms,and the algorithm efficiently reduces runtime.The result of experiment indicates that the algorithm is more fast and efficient than presented algorithms in spatial data mining.