摘要
基于模糊C-均值(FCM)算法中类中距与类间距的含义设计了一个新的模糊分类适应度函数。再分别以本文设计的适应度函数及国际上常用的几种分类准则函数转换成的适应度函数,利用遗传—迭代自组织分析技术(GA-ISODATA)共同执行模糊C-均值(FCM)的优化计算,并对结果进行分析比较,最后给出设计最佳的适应度函数应遵循的原则。实例研究表明,该适应度函数适合社会经济系统软分类。
This paper firstly designs a new fuzzy classified fitness function based on inner-distance and inter-distance of fuzzy C-means （FCM）; then separately performs FCM according to the fitness function designed by author and the fitness function transformed from several international classify rulers, and compares the result; at last gives some principles that designing optimal fitness function must follows. The example study indicates that the fitness function designed by author is especially fit for soft classify of social and economic system.
出处
《系统工程理论方法应用》
北大核心
2006年第3期241-246,255,共7页
Systems Engineering Theory·Methodology·Applications
基金
国家自然科学基金资助项目(70273044)
关键词
遗传算法
适应度函数
最佳适应度函数
设计原则
genetic arithmetic
fitness function
optimal fitness function
design principle
