论文部分内容阅读
近来一些学者用差别矩阵或差别矩阵的思想设计了基于正区域的属性约简算法.由于计算差别矩阵是一个既消耗时间又消耗空间的过程,故这些算法的效率并不好.为了降低这类属性约简算法的复杂度,文中利用基于区分对象对的属性约简的思想,在简化决策表的基础上,定义了一个函数,该函数能度量简化决策表中条件属性集产生的区分对象对的个数,并用该函数设计了一个启发函数,同时给出了计算该启发函数的快速算法,经分析其时间和空间复杂度均为O(|U/C|).最后用该启发函数设计了一个有效的基于正区域的属性约简算法,该算法的时间复杂度降为O(|C||U|),空间复杂度降为O(|U|).文中还用一个具体实例说明了新算法的有效性.经实验证明,新算法具有较高的效率.
Recently, some scholars have designed the attribute reduction algorithm based on the positive region based on the idea of differential matrix or differential matrix. Since the computational difference matrix is a time consuming and space consuming process, the efficiency of these algorithms is not good. In this paper, based on the idea of attribute reduction based on object-oriented pairs, a function is defined based on the simplified decision table. This function can measure and simplify the pairs of distinguishing objects generated by the set of condition attributes in the decision table , And a heuristic function is designed by this function. At the same time, a fast algorithm for calculating this heuristic function is given, and its time and space complexity are both O (| U / C |). Finally, this heuristic function design An effective attribute reduction algorithm based on the positive region is proposed, whose time complexity is reduced to O (| C || U |) and the space complexity is reduced to O (| U |) .There is also a concrete example The validity of the new algorithm.Experimental results show that the new algorithm has higher efficiency.