一种基于蚁群优化的粗糙集属性约简方法

Method of Attribution Reduction of Rough Set Theory Based on MMAS

针对基于智能优化算法的属性约简方法存在的问题，在扩张矩阵概念的基础上，将属性约简问题转化为几何路径寻优问题，通过改进最大—最小蚂蚁系统（MMAS）算法，提出了一种基于蚁群优化的粗糙集属性约简方法，给出了相关参数的定义和详细算法流程，在UCI数据集上测试了该算法，结果表明了算法的有效性。

Aiming at problems of attribution reduction, on the basis of dilation matrix, the problem of reduction was translated into the geometrical optimization problem. The method of attribution reduction of rough set theory based on MMAS was put forward. Correlative parameters were defined and the algorithm flow was presented in this paper. Factors, which affected the algorithm efficiency, were analyzed. The test result proved the validity of the method.