多层关联规则挖掘算法的研究及应用

Research and application of multi-association rule mining algorithm

针对商业银行业务系统中海量数据的分析和研究问题,提出了一种改进频繁项集挖掘算法FP-growth的多层关联规则数据挖掘算法。在对大量商业银行业务交易处理内在规律研究的基础上,依据利润度进行划分,使得该算法在满足用户需求的基础上,有效的缩小了层次结构树的规模,又加快了搜索的速度,从而提高了数据挖掘的效率。模拟算例表明,该算法有效可行,能够更好地适应商业银行交易系统层次结构在大型数据集的数据挖掘。

Aiming at the problem of a large number of transactions in a banking business systems,an improved multilevel association rules algorithm based on the FP-growth algorithm of data mining which improves the frequent itemsets of the data mining is put forward.By analyzing the massive data of the current operational system,the algorithm is classfied based on division of profits.The size of the tree is significantly reduced,and on the basis of the tree,a maximal target frequent itemset mining algorithm is put forward,which satisfies the users requirements and accelerates the speed to traverse the tree,so the mining efficiency is improved in the algorithm.A simulation example is presented to prove the proposed method.Consequently,the proposed algorithm can better adapt to the hierarchical structure of the commercial banking system in large-scale data sets data mining.