一种基于遗传算法的分裂式层次化聚类算法

A GA-based divisive hierarchical clustering algorithm

针对聚类中自适应确定聚类个数、目标函数灵活定义及优化的近似计算等问题,综合了分裂式层次化聚类算法能根据相似度阈值自适应地确定聚类个数的特点及二进制遗传聚类算法具有较强的搜索近似最优解能力及目标函数定义灵活的特点,提出了一种基于遗传算法的分裂式层次化聚类方法。实验结果表明,该算法具有较好的聚类性能。

To solve the problems of adaptive determinition of the cluster number, flexible objective function definition and approximate optimal computation in clustering analysis, a GA-based divisive hierarchical clustering algorithm （GADHC） was proposed, which integrated some features of divisive hierarchical clustering algorithm and binary genetic clustering algorithm. The experiments show that the proposed algorithm works well.