基于邻域粒化混合决策系统克隆选择约简算法

Reduction algorithm of hybrid decision system based on neighborhood granulation and niche clone selection

针对混合决策系统的属性约简问题,提出了基于邻域粗集模型的小生境克隆选择属性约简算法.采用邻域关系度量不可分辨关系,通过邻域信息粒子逼近论域空间,可以直接处理数值型属性.克隆选择约简算法的提出解决了求解全部约简的NP完全问题.论述了亲和度函数的选择,引入了小生境技术,避免了抗体的早熟收敛及算法中的参数对具体优化目标的敏感性和单一收敛性,给出了算法的具体实现.对经典数据集和UC I中4组数据约简的仿真结果证明了算法的有效性和可行性.

In order to reduce the hybrid decision system,a reduction algorithm is proposed based on the neighborhood rough set model and niche clone selection algorithm.In the model the indiscernibility relation is measured by neighborhood relation,and the universe spaces is approximated by neighborhood information granules,so the numerical attributes can be treated directly.The fitness function is designed,and the reduction algorithm is presented as well.The introduction of niche technology can avoid the early convergence of the antibody,and can avoid the sensitization and the local astringency of the parameter to the specific optimal objects.The validity and feasibility of the algorithm are demonstrated by the results of experiments on a classical data set and four UCI machine learning databases.