摘要
研究了基于依赖度的属性约简算法,并以实例说明该算法不适合于不相容的决策表,在分析算法产生缺陷原因的同时通过改变算法求核的终止条件,使得算法能够同时适用于相容和不相容的决策表.在求约简过程当中,首先对除核之外剩余属性重要度进行计算,然后以重要度大小依次添加入核求约简.最后通过实例来验证改进后算法的有效性.
The algorithm based on the dependability of attribute reduction is discussed in this paper and proved to be unsuitable for the incompatible decision table. After analyzing the defects of this algorithm, this paper proposes an improved one which is suitable for the compatible and incompatible decision table through changing the termina- tion condition of the core. In the reduction process, it calculates the importance of the remaining attribute excepting the core, and then to add to the size of an important degree in the core for reduction. At last, an example is used to prove the validity of this improved algorithm.
出处
《云南民族大学学报（自然科学版）》
CAS
2011年第6期504-507,共4页
Journal of Yunnan Minzu University:Natural Sciences Edition
基金
云南省教育厅科学研究基金(2010J071
09Y0258)
关键词
粗糙集
属性约简
依赖度
rough set
attribute reduction
dependability
