最优特征子集的遗传算法求解

Optimal Feature Subset Selection Using Genetic Algorithms

指出最优特征子集选择问题(OFSS)是个NP-Hard问题,寻找一个近似算法具有现实意义。遗传算法提供了一种求解复杂系统优化问题的通用框架。使用基于小生境技术的遗传算法求解OFSS问题,以获得较好收敛性、稳定性和较快的速度。

As optimal feature subset selection （OFSS） is an NP-hard Problem, it is of realistic significance to find an approximative algorithm. Genetic algorithms （GA） provide an all-purpose frame work to solve complicate optimization problem. To solve the problem of OFSS, this paper brings forward an algorithm based on niched genetic algorithms （NGA）, which is of rather better astringent and stability.