论文部分内容阅读
目前设计基于差别矩阵的求核算法的主要方法是差别矩阵方法.在该种方法中,是通过搜索差别矩阵的所有差别元素得到核.由于是在所有的差别元素上搜索,故该方法比较耗时.本文在简化决策表和简化差别矩阵的基础上,将具有核属性的差别元素集归纳在某一相对较小的集合上,故新算法只需搜索和检查简化差别矩阵的少量差别元素就可以得到核算属性集.设计了一个高效求核算法,其时间复杂度为max{O(|C|2|U/C|),O(|C||U|)},其空间复杂度为O(|U|).由于新算法只判断简化差别矩阵的少量差别元素就可以找到核算属性集,故新算法的效率得到了有效地改善.
At present, the main method of kernel algorithm based on discernibility matrix is the discernibility matrix method, in which the kernels are obtained by searching all the difference elements of the discernibility matrix. Since it is searched on all the difference elements, , This paper simplifies the decision table and simplifies the discernibility matrix based on the nuclear attributes of the differential elements are summarized in a relatively small set, so the new algorithm simply search and check the simplified differential matrix of a small number of different elements We can get the set of accounting attributes.An efficient computational algorithm is designed, whose time complexity is max {O (| C | 2 | U / C |), O O (| U |). The efficiency of the new algorithm is effectively improved because the new algorithm can find the set of accounting attributes only by judging a few difference elements of the simplified difference matrix.