基于十字链表的Separate算法研究  

Research of Separate Algorithm Based on Cross Linker

作　　者：陈平[1] 王利钢[1] 

机构地区：[1]南京信息职业技术学院信息服务学院,南京210023

出　　处：《信息化研究》2015年第5期28-33,共6页INFORMATIZATION RESEARCH

摘　　要：生成频繁项集是约束关联规则挖掘中最耗时的一步。为提高挖掘的效率,文章提出一种基于十字链表的经典的约束算法(Separate)改进算法(SABCL)。该算法结合Separate算法和十字链表存储的优势,先根据约束条件过滤数据库并产生频繁项目,再利用十字链表存储频繁项目信息,通过遍历十字链表生成满足约束条件的频繁项集,从而避免重复扫描数据库,提高了挖掘效率。在美国蘑菇(mushroom)、象棋(chess)数据集上,对SABCL、Separate算法进行了实验对比,实验结果表明,在相同的条件下,SABCL算法能够快速地生成所有的频繁项集,时间性能优于Separate算法。Generating frequent item sets in constraint association rules mining is the most time consuming step.To improve the efficiency of mining,an improved algorithm named separate based on across linker(SABCL)is proposed.The algorithm combines the advantages of the algorithm named Separate and across linker in the storage.Firstly,database is filtered and frequent item is generated by the constraint condition.Then the frequent item's information is stored by across linker and the frequent item sets that meet constraints are generated by traversing the across linker to avoid scanning database repeatedly and improve the mining efficiency.Finally,We compared the run time of SABCL,Seperate algorithm in mushroom and chess datase.The experimental results indicate that SABCL can quickly generate all frequent item sets,and has better time performance than Separate algorithm in the same conditions.

关 键 词：关联规则 十字链表 项约束 频繁项目集 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]