【摘要】 粗糙集理论是应用于知识约简及分类等领域的一种高效数学工具。基于欺诈识别中的布尔型决策表,提出了属性幂的概念,据此可以快速计算划分,缩短运算时间。应用粗糙集相关理论,并结合排序思想,提出一种新的决策表属性约简方法。实例计算及应用于信用数据上的实验验证了算法的可行性及有效性。更多还原

【Abstract】 Rough set theory is an efficient mathematical tool applying on domains of knowledge reduction and classification etc.Basis on the Boolean decision table in fraud detection,the concept of attribute power is proposed so as to compute partition quickly and reduce the algorithm time.With the related theories of rough sets and using the ordering idea,this paper puts forward a new method of attribute reduction.The algorithm is proved feasible and efficient through a real example and the experiment on a credit data set.更多还原