单调和反单调约束条件下关联规则的挖掘算法分析.pdfVIP

计算机科学2005Voi．32N9．6 单调和反单调约束条件下关联规则的挖掘算法分析¨ 杜剑峰李宏陈松乔陈建二 (中南大学信息科学与工程学院长沙410083) 摘要本文充分利用了Eclat算法的概念格理论和等价类划分方法，将约束冬件融入基于垂直数据分布的关联规 算法采用自底向上的搜索方法，在发现频繁项集的同时进行约束条件的检验。数据库的扫描次数较少，无需对候选项 集进行剪枝，占用内存较小。实验证明：该算法的执行效率比已有算法有显著提高。 关键词数据挖掘，约束性关联规则，概念格，等价类 forAnti-monotoneandMonotoneConstraintsAssociationRules AlgorithmAnalysis Mining CHEN DU LI CHEN Jian。-Er Jian——FengHong Song—-Qiao South 410083) ofInformation University，Changsha (College Engineering，Central are intothe ofassociationrules AbstractAnti—monotoneandmonotoneconstraints integratedalgorithm mining basedOnverticaldata thelattices andthe methodof classessuffi— layoutbyutilizing theory decomposingequivalence is anti—monotoneconstraintnamedEclatA．Theother new arealso basedon ciently．Twoalgorithmspresented．One isbasedon constraintnamedEclatM．A searchmethodis forwardandtheconstraintsare monotone bottom～up put fewtimesandhaveno checkedatthe of itemsets．Thenew scanthedatabase processcalculatingfrequent algorithms need candidateitemsets．Atthesame caD-besolvedin fromour of experiments pruning time，they memory．Results showthatthe ofthenew isunmatched performance algorithms byanypreviousalgorithms Dat