A Bottom-Up Enumeration Algorithm of Minimal Generators without Support Counting for Compressing Negative Association Rules
A Bottom-Up Enumeration Algorithm of Minimal Generators without Support Counting for Compressing Negative Association Rules
复制标题
DOI：
10.1109/iiaiaai55812.2022.00135
复制
发表时间：
2022-07
期刊：
影响因子：
--
通讯作者：
Kento Yajima;K. Iwanuma;Yoshitaka Yamamoto
中科院分区：
文献类型：
--
作者：
Kento Yajima;K. Iwanuma;Yoshitaka Yamamoto
负关联规则对于表达大数据中隐藏的各种潜在属性非常有价值且必不可少。然而，有效的负关联规则的集合总是变得如此巨大，因此有效的负规则压缩方法就显得非常重要。最小生成器是压缩有效否定规则集的重要概念。在本文中，我们研究了一种自下而上的有效算法，用于从给定的封闭项集中枚举最小生成器。到目前为止，支持计算一直是数据挖掘计算的瓶颈。因此，我们在这里提出了一种新颖的枚举算法，该算法不使用任何显式支持计算。我们展示了评估自下而上枚举算法的实验结果，与以前的算法相比，该算法在内存消耗方面表现良好。
Negative association rules are valuable and essential for expressing various latent properties which hide in big data. The set of valid negative association rules, however, always becomes so huge, thus an effective compression method of negative rules is quite important. Minimal generators are important concepts for compressing the set of valid negative rules. In this paper, we study a bottom-up efficient algorithms for enumerating minimal generators from given closed itemsets. So far, support calculations have always been a bottleneck in data mining calculations. Therefore, we here propose a novel enumeration algorithm which does not use any explicit support computation. We show experimental results for evaluating our bottom-up enumeration algorithm, where the algorithm performs well in terms of memory consumption compared to previous algorithms.