摘要
为解决在挖掘频繁项集时由忽略项目间重要性差异以及最小支持度频繁变动而导致的挖掘效率低以及利用率低。通过关系矩阵解决数据体量大造成的挖掘效率低的问题;通过加权规则解决不同业务项目间重要性差异问题;通过动态树解决最小支持度变动频繁的问题。本文创新性提出加权矩阵动态树算法WMDT。实验结果表明,WMDT算法较以往算法,精准度和挖掘效率有显著提高同时受最小支持度变动影响较小,是一个高效的频繁项集挖掘算法。
In order to solve the problem of mining frequent itemsets,the mining efficiency is low and the utilization rate is low,which is caused by ignoring the importance difference between items and the frequent change of minimum support. The problem of low mining efficiency caused by large volume of data is solved by relation matrix,the problem of importance difference between different business items is solved by weighting rules,and the problem of frequent change of minimum support degree is solved by dynamic tree. This paper proposes a weighted matrix dynamic tree algorithm WMDT. Experimental results show that WMDT algorithm is an efficient frequent itemsets mining algorithm,which can significantly improve the accuracy and mining efficiency compared with previous algorithms,and is less affected by the change of minimum support.
作者
杨秋翔
王冠男
王婷
YANG Qiu-xiang;WANG Guan-nan;WANG Ting(School of Software,North University of China,Taiyuan 030051,China)
出处
《电子设计工程》
2019年第15期66-70,75,共6页
Electronic Design Engineering
关键词
频繁项集
关系矩阵
加权规则
动态树
frequent itemsets
relation matrix
weighted rule
dynamic tree
