一种基于关系积的属性值约简算法

Value reduction algorithm based on attribute union

值约简是粗糙集(Rough Set,RS)理论的一个重要研究课题,但由于值约简是一个NP-hard问题,目前还没有一个高效的值约简算法。基于集合理论,提出了关系积概念,把决策表的属性约简过程转化为关系积的运算,提高了属性约简的效率;利用各阶关系积生成时的信息,获得最小值约简表,从而解决了值约简这一NP问题。

The value reduction is an important research topic in Rough Se（tRS） theory.Because value reduction is a NP-hard problem,there is no effective algorithm to get the minimal value reduction at present.The attribute reduced procedure is translated to find the attribute union based on the set theory,reducing the number of scanning the decision table and improving the reduced efficiency;Using the information derived from the attribute reduction algorithm,the minima value reduction table is obtained and this NP-hard problem is solved.