一个基于差别矩阵思想的高效求核算法

An Efficient Algorithm for Counting Core Based on Discernibility Ma trix Thinking

目前,关于属性约简已有不少算法,其中在很多算法中,都要求先求出核属性集,但利用差别矩阵求核属性这一算法中,生成差别矩阵时,有许多不必要的元素被生成,这些无用的元素在求核时又要进行比较,因而效率不高。利用差别矩阵的思想设计一种不必生成那些不必要的元素的求核算法,从而使算法的效率得到提高。最后,给出了一个实例说明新算法的高效性。

At present ,there are many algorithms about attribute reduction,but in these algorithms ,it is necessary that the set of core attributes are get in using discernibility matrix at first.Because there are many unused elements in the created discernibility matrix,these unused elements are compared in counting core,the efficient of these algorithms for counting core which use discernibility matrix is not good.In this paper,we give an algorithm which use discernibility matrix thinking,but in this algorithm it is unnecessary that create these unused elements in discernibility matrix,so the new algorithm is more efficient than the old one.At the end,we give an example which explain the efficient of the new algorithm.