一种基于频率的多最小支持度挖掘算法  

A Multi-minimum Support Mining Algorithm Based on Frequency

作　　者：古良云 乐红兵 GU Liangyun;LE Hongbing(Jiangnan University,Wuxi 214122)

机构地区：[1]江南大学,无锡214122

出　　处：《计算机与数字工程》2020年第12期2942-2947,共6页Computer & Digital Engineering

摘　　要：关联规则是描述数据集中不同项之间的稳藏关系,现有的挖掘算法大多是挖掘数据集中满足用户指定的最小支持度和最小置信度约束的所有关联项。由于仅有一个最小支持度用于整个数据集,如果最小支持度很高,则不能发现出现频率较低的项目;如果最小支持度太低,则会出现规则爆炸的现象。为此,提出一种基于频率的多最小支持度挖掘算法,该算法将数据集中各项目实际频率作为其最小项目支持度,并通过设置项目频率最小阈值控制频繁项集的产生。在合成数据集、Zoo数据集和kaggle提供的数据集上的实验结果表明,该算法能更加有效地挖掘关联规则。The association rule describes the stable relationship between different items founded in data sets.The existing min⁃ing algorithms are all related items in the data set that satisfy the user-specified minimum support and minimum confidence con⁃straints.Since only one minimum support is used for the whole data set.If the minimum support is too high,it cannot find a project with a lower frequency,if the minimum support is too low,a rule explosion will occur.Therefore,a multi-minimum support mining algorithm based on frequency is proposed,which takes the actual frequency of each item in the data set as its minimum project sup⁃port,and controls the generation of frequent itemsets by setting the project frequency minimum threshold.Experimental results on synthetic datasets,Zoo datasets and datasets provided by Kaggle show that the algorithm can more effectively mine association rules.

关 键 词：关联规则 频率 多最小支持度 项目频率最小阈值 数据挖掘 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]