一种结合人工蜂群和K-均值的混合聚类算法

传统的K-均值聚类算法虽然收敛速度快,但由于过度依赖初始聚类中心,算法的鲁棒性较差。为此,提出了一种改进人工蜂群算法与K-均值相结合的混合聚类方法,将改进人工蜂群算法能调节全局寻优能力与局部寻优能力的优点与K-均值算法收敛速度快的优点相结合,来提高算法的鲁棒性。实验表明,该算法不仅克服了传统K-均值聚类算法稳定性差的缺点,而且聚类效果也有了明显改善。

Hybrid clustering algorithm based on artificial bee colony and K-means algorithm

The traditional K-means clustering algorithm is too dependent on the initial clustering centers. With regards to this, this paper proposed a mixed clustering method based on the improvement artificial colony algorithm and the K-means algorithm. The new method combined the advantages of regulating ability of global optimization and local optimization with rapid convergence of K-means clustering algorithm to improve the robustness of the algorithm. Experiments show that the clustering result of the new method is significantly improved, not only the stability.