基于混合核函数的可能性C-均值聚类算法

针对传统的模糊C-均值算法对于非球形分布的数据聚类效果不理想且易受到噪声数据的影响,利用可能性C-均值算法具有良好的抗噪声性能,将混合核函数引入到该算法中,提出了一种基于混合核函数的可能性C-均值(HKPCM)聚类算法。该算法将原空间的待分类样本映射到一个高维的特征空间(核空间)中,使得样本变得线性可分,然后在核空间中进行聚类。实验结果证实了HKPCM算法的可行性和有效性。

Possibilistic C-means clustering algorithm based on hybrid kernel function

Traditional fuzzy C-means algorithm have a bad clustering result for non-spherical data and it is easy to be affected by the noise data. To solve these problems, this paper used the advantage of possibilistic C-means clustering algorithm and combined hybrid kernel function, proposed HKPCM algorithm. The samples in the original space were mapped into a high dimensional space by using this algorithm. So the samples became linearly separable and it was easy to cluster in kernel space. Expe-riment results indicate that the HKPCM algorithm is feasible and efficient.