基于遗传算法的一种自动聚类方法

遗传算法(Genetic Algorithm,简称GA)是一类模拟生物界的进化规律(适者生存,优胜劣汰遗传机制)而形成的一种最适应全局优化概率搜索算法。针对常规动态聚类方法对初始聚类中心的敏感性以及聚类结果与样本输入次序有关等问题,本文提出了一种基于GA的动态聚类方法,并将它应用到数据库的数据分析中。计算结果表明,该方法是一个具有全局最优解的动态聚类方法,其结果明显好于K-均值聚类算法。

A Dynamic Clustering Method Based on Genetic Algorithms

Genetic Algorithms (GA), which simulates the evolution laws (mechanism of being survival of the fittest) of biological world, is an algorithm with global optimization. To solve the problem of sensitivity with the original clustering center and clustering results depended on the order of the input example in common dynamic clustering algorithm, a new dynamic clustering method based on genetic algorithms is presented in this paper and is applied to data analysis in databases. Computing results indicate that the method is a dynamic clustering algorithm with global optimization and is superior to K- means algorithm.