对于某些模糊聚类算法的比较研究

    A Comparative Research on Some Fuzzy Clustering Algorithms

    • 摘要: 本文介绍了模糊C均值聚类算法,诱导的模糊C分划算法,基于非对称不相似性的系统聚类算法以及模糊集的峰值搜索算法,给出了关于几个实验数据集的聚类结果,并讨论了这些算法的性质。结果表明,系统聚类算法在这四种算法中最有效,而峰值搜索法优于模糊C均值和诱导的模糊C分划算法。

       

      Abstract: In this paper, the fuzzy C-means clustering algorithm(FCM), the induced fuzzy C-partition algorithm(IFCM), the hierarchical clustering algorithm based on asymmetric similarities(HCAM), and the algorithm for detecting unimodal fuzzy sets(GFAM) are described. The clustering results of several experimental data sets are given. The properties of these algorithms are also discussed. The results obtained indicate that HCAM is the most powerful algorithm of the four, and that GFAM is more powerful than FCM and IFCM,

       

    /

    返回文章
    返回