基于概念格的无冗余关联规则提取算法  

CONCEPT LATTICE-BASED EXTRACTION ALGORITHM FOR NON-REDUNDANCY ASSOCIATION RULES

作　　者：翟悦[1] 秦放[1] 

机构地区：[1]大连科技学院信息科学系,辽宁大连116052

出　　处：《计算机应用与软件》2015年第4期46-49,66,共5页Computer Applications and Software

摘　　要：针对传统挖掘算法生成的关联规则存在大量冗余、难于理解和应用的问题,提出一种新的频繁闭项集概念格FCIL(Frequent Closed Itemsets Lattices),用于生成无冗余关联规则。首先,对概念格理论进行研究,概念格节点间的泛化和例化关系非常适合规则提取;然后,结合频繁闭项集能有效减少规则数目的特点,构建一种新的FCIL;最后,给出FCIL构造算法和相应的规则提取算法。实验表明,该方法能够高效地产生无冗余规则集。Association rules generated by traditional mining algorithm have a large number of redundancies,which are hard to be understood and applied.Aiming at this problem,we present a new frequent closed itemset concept lattice （FCIL）,which can generate non-redundancy association rules.First,we study the theory of concept lattice,the generalisation and specialisation relations between the nodes of concept lattice are very suitable for rule extraction;then,we build a new FCIL in combination with the characteristic of frequent closed itemsets which can effectively reduce the number of rules;finally,we develop the FCIL construction algorithm and the corresponding rules extraction algorithm.Experiment shows that the proposed method can efficiently generate non-redundancy rules set.

关 键 词：频繁闭项集 FCIL 无冗余关联规则 

分 类 号：TP301.06[自动化与计算机技术—计算机系统结构]