基于遗传算法的不完备信息系统属性约简方法

提出了一种基于遗传算法的属性相对约简算法。通过在不完备信息系统中引入决策属性支持度的概念,来描述由条件属性所提供的知识对整体决策的支持程度,以此作为启发式信息求出相对核,并将相对核加入遗传算法的初始种群中以加快算法的收敛。同时,在适应值函数中引入惩罚函数,可以保证所求约简既含较少的属性又有较强的支持度,能够获得最佳的搜索效果。该算法通过实例分析,证明是求解属性约简问题的快速有效方法。

Attribute reduction based on GA under incomplete information system

A kind of attribute relative reduction algorithm is proposed.With decision attribute support degree applied in a incomplete information system,the support degree of the knowledge supplied by condition attribute for the whole decision is described and the relative importance degree and relative core are obtained as initial population in GA in order to accelerate convergence.A punishing function is led into the fitness function to assuring reduction having fewer attributes and stronger support and search effect is very good.The practical results show that the approach is effective in solving attribute reduction.