摘要
针对不相容决策表中一些属性约简算法的不足,结合粗糙集的代数观与信息观的优点,对差别矩阵加以改进,提出了一种新的属性约简算法,该算法在保证约简后决策表的正域和条件信息熵不变的情况下,降低了时间复杂度。通过实例说明了该算法的有效性和可行性。
In order to solve the disadvantages of the attribute reduction algorithm for inconsistent decision table,combining information view and the algebra view in rough set,the discernibility matrix is improved,a new method of attribute reduction algorithm is presented,which reduces the time complexity in the case of the positive regions and information entropy unchanged after attribute reduction.The example of computing and analyzing with the proposed method shows its advantages and feasibility.
出处
《计算机工程与应用》
CSCD
北大核心
2011年第2期124-126,共3页
Computer Engineering and Applications
基金
国家自然科学基金(No.50675069)~~
关键词
粗糙集
差别矩阵
属性约简
rough set
discernibility matrix
attribute reduction
