摘要
目前,对布尔型关联规则的挖掘研究已较成熟,而对量化关联规则的挖掘研究相对较少,并且采用的挖掘方法多是将量化属性进行离散化处理,进而转化为布尔型关联规则进行挖掘。但传统的对量化属性离散化处理的方法存在区间划分过硬的问题,因此提出一种基于数据场的量化关联规则挖掘方法。该方法避免了区间划分过硬问题,同时也充分考虑了数据集中数据的非完备性以及每个数据对数据挖掘任务所发挥的不同作用。实验证实了该方法的有效性。
At present, the research on Boolean association rule mining has been quite mature, while the research on quantitative association rule mining is relatively rarer, and most of the mining methods used are to carry out diseretisation treatment on quantify properties, then further transform the quantitative association rule to Boolean association rule for mining. However, traditional discretisation processing method of quantitative attributes has the problem of tough interval division. Therefore, we present a data field-based quantitative association rules mining method, it prevent the tough interval division problem, and at the same time fully considers the incompletion of data in dataset and the different roles of each data in data mining task. The validity of this method is testified by the experiments.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第7期40-42,58,共4页
Computer Applications and Software
基金
内蒙古高等教育科学重点项目(NJZZ11140)
内蒙古自然科学基金项目(2012MS0611)
关键词
数据挖掘
量化关联规则
数据场
聚类分析
Data mining
Quantitative association rules
Data field
Cluster analysis
