摘要
在地震数据挖掘应用中,可用粗糙集方法进行对震例数据的属性约减。但是,经典的粗糙集理论建立在由等价关系对对象集划分的基础上,而震例数据是有序的而不是分类的对象。现对经典粗糙集理论进行扩展,提出一种用优势关系代替等价关系的粗糙集方法,并在此基础上提出一种基于差别矩阵的属性约简算法。实验结果表明,用这种方法能得出一些采用传统粗糙集理论所无法得到的结果。
In seismology, rough set can be used in the reduction of attributes. But classical rough set are based on indisceroibility relation or similarity relation. The problems in Seismology deal with ordering objects instead of classifying objects. Now we presented an extension of the classical rough set called dominance-based rough set, and an attribute-reduction algorithm based on dominance matrix. The experiment indicates that the results will be more meaningful than those induced by the classical rough set theory.
出处
《计算机应用》
CSCD
北大核心
2006年第12期3023-3026,共4页
journal of Computer Applications
基金
国家地震科学联合基金资助项目(104090)
上海市自然科学基金资助项目(7A05468)
