一种基于粒度的规则挖掘方法  

A rules mining method based on granularity

作　　者：顿毅杰[1] 张小峰[1] 孙昊[1] 赵丽[1] 

机构地区：[1]兰州理工大学计算机与通信学院

出　　处：《兰州理工大学学报》2006年第1期105-108,共4页Journal of Lanzhou University of Technology

基　　金：甘肃省自然科学基金(ZS022-A25-035)

摘　　要：利用知识的粒度,提出了一种新的规则挖掘方法.首先利用粗糙集中判断属性重要性的尺度,获取一个属性,由此属性形成对域的粒度,并对粒度中的每一个粒进行同样操作,直至所有粒的决策属性值唯一为止.算法分析与实验对比说明,该方法是有效可靠的.The knowledge granularity was used to present a new method to mine rules based on granularity. First, a measure for judging the importance of attribute in a rough set was used to obtain an attribute, and then get the granularity of the universe. And then this procedure will be repeated for every granule of the granularity until the decision attribute of all granules becomes a unique value. The analysis of the algorithm and the experiment on it shows that the method is effective and reliable.

关 键 词：粒 粒度 分类规则 知识发现 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]