将分治法的思想溶人Rough集算法中,在给定属性序下,提出了基于分治策略的属性约简算法.利用该算法可以计算给定属性序下的唯一约简,并能快速得到海量数据的属性约简.在一次性将决策表的所有数据调入计算机内存的情况下,算法的平均时间复杂度为O(|U|×|C|×(|C|+log|U|)),空间复杂度为O(|U|+|C|).仿真实验结果说明了算法的高效性.
The idea of divide and conquer is adopted in attribute reduction of rough set theory. A quick algorithm for attribute reduction of ordered conquer method. A unique attribute reduction can for dealing with huge data reduction. If all data of attributes is proposed based on the divide and be obtained with this algorithm. It is suitable a decision table could be loaded in memory one time, the average time complexity of this algorithm will be O(|U|×|C|×(|C|+log|U|)) and its space complexity will be O(|U|+|C|). Simulation experimental results show its efficiency.