一种基于矩阵的频繁项集更新算法

Updating algorithm based on matrix for mining frequent item sets

针对相关算法在处理频繁项集更新时所存在的问题,提出了一种基于矩阵的频繁项集更新算法。该算法首先以时间为基准将更新后的数据库分为原数据库和新增数据库,分别将它们转换为0-1矩阵,通过矩阵裁剪、位运算产生新增频繁项集,并利用已有频繁项集更新原有频繁项集。实验仿真结果不但证明了该算法的可行性和高效性,而且还证明了它适合大型、稠密性数据库的频繁项集更新。

Aiming at updating problems of frequent item sets, this paper proposed an updating algorithm based on matrix （UABM） for mining frequent item sets. Divided the updated database into original database and new one based on time. Convetted these databases into matrixes. Got the new frequent sets by matrix cropping and the bit operation, and updated the gotten frequent item sets on gotten ones. The experiments show the algorithm is not only feasible and efficient but also fit to update freguent item sets for a large-scale and dense data base.