基于矩阵的不确定数据频繁项集快速挖掘算法  

Fast algorithm of frequent itemset mining based on matrix from uncertain data

作　　者：刘芝怡[1] 常睿[2] 

机构地区：[1]常州工学院计算机信息工程学院,江苏常州213002 [2]常州工学院计划财务处,江苏常州213002

出　　处：《南京理工大学学报》2015年第4期420-425,共6页Journal of Nanjing University of Science and Technology

基　　金：江苏省自然科学基金(BK20130245)

摘　　要：针对CUF-growth算法中项集的期望支持度估算值过大,且挖掘过程中需要反复递归构造条件CUF-tree导致挖掘效率降低这一问题,提出UFIM-Matrix(Uncertain frequent itemset mining-matrix)算法。该算法不需要建立树结构,而是利用计算项集估算期望支持度的新方法和矩阵结构来产生规模更小候选项集,能在一定程度上减少计算开销,提高挖掘效率。最后的实验结果也表明了新算法性能更优。The CUF-growth algorithm gives an upper bound on the expected support of itemsets,but the estimate is too high. It has own bottleneck that needs to build conditional CUF-tree recursively in the mining process for getting candidate itemsets. According to the deficiency of the CUF-growth,the UFIM-Matrix( Uncertain frequent itemset mining-matrix) algorithm is proposed. This algorithm does not need to build a pattern tree while it generates smaller candidate sets by using a matrix structure and an improved method to calculate the upper bound of the expected support of itemsets. It can greatly reduce the cost of computing and improve the mining efficiency. The experimental results indicate the algorithm is more effective and efficient.

关 键 词：不确定数据 频繁项集 期望支持度 快速挖掘 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]