Exceptional Association Rule Set Discovery from Community-Dwelling Elderly People Database
Exceptional Association Rule Set Discovery from Community-Dwelling Elderly People Database
复制标题
DOI：
10.1109/smc.2018.00558
复制
发表时间：
2018-10
期刊：
影响因子：
--
通讯作者：
K. Shimada;H. Aoki;K. Kubota;S. Haresaku;S. Mizutani;T. Naito;Michio Ueno
中科院分区：
文献类型：
--
作者：
K. Shimada;H. Aoki;K. Kubota;S. Haresaku;S. Mizutani;T. Naito;Michio Ueno
提出了一种从不完备数据库中发现异常关联规则集的扩展方法。在一个例外规则集中，每个项目集X，Y与类C有弱的统计关系或没有统计关系;然而，X和Y的连接与C有强的关系。异常规则集可以用来推断X和Y的连接的长规则，并发现罕见的规则。该方法直接计算规则评价优势比。在这项研究中，该方法被应用到规则发现使用数据库的社区居住的老年人。实验结果表明，该方法可以帮助发现有趣的稀有规则和例外的关联规则集。实验结果表明了该方法在医疗卫生领域的有效性。
An extended method to discover exceptional association rule sets from incomplete databases is proposed. In an exceptional rule set, each itemset X, Y has a weak or no statistical relation to class C; however, the join of X and Y has a strong relation to C. An exceptional rule set can be used to infer long rules for the join of X and Y and to discover rare rules. The proposed method calculates the rule evaluation odds ratio directly. In this study, the method is applied to rule discovery using a database of community-dwelling elderly. Experimental results demonstrate that the proposed method can help discover interesting rare rules and exceptional association rule sets. The results show the effectiveness of the proposed method in the fields of medicine and health care.