摘要
提出了一种用于聚类分析的克隆-K均值算法。基于人工免疫系统的克隆选择算法具有全局搜索能力强,收敛于全局最优解的特点。基于以上优点,在克隆选择算法中引入K-均值算子,对种群中的个体在克隆、变异操作后进行K-均值运算。通过对初始种群的形成、克隆操作、变异操作、替代操作和K-均值操作等过程的描述,提出了完整的克隆-K均值算法。实验研究表明,算法成功解决了K-均值算法对初始值敏感且容易陷入局部最优的缺点,算法明显优于传统的K-均值聚类算法。
A Cloning - K - means Algorithm for cluster analysis is given. The Colonel Selection algorithm based on artificial immune system has strong global search capability and converges to the global optimal solution. In this paper, the K - means operator is added to the Colonel Selection Algorithm and it' s used for the individual of the whole group after the individual has been cloned and mutated. From the description of the generation of initial group, the cloning operation, the Mutation operation, the Alternative operation and the K - means operation, an intergraded Cloning - K - means Algorithm is presented. The experiment demonstrates that the problem of the K - means algorithm has been solved and the Cloning - K - means Algorithm is superior to the pure K - means algorithm.
出处
《计算机仿真》
CSCD
2008年第11期191-194,共4页
Computer Simulation