一种快速的广义噪声聚类算法

Fast generalized noise clustering algorithm

为解决广义噪声聚类(GNC)算法非常依赖参数和在运行GNC算法前必须运行FCM算法以便计算参数的缺点,在GNC的目标函数和可能聚类算法(PCA)基础上,提出一种快速的广义噪声聚类(FGNC)算法。FGNC算法通过一种非参数化方法计算GNC目标函数中的参数,因而FGNC算法不依赖参数并且聚类速度快于GNC算法。对人工含噪声数据集和两个实际数据集进行仿真实验,实验结果表明FGNC算法能很好地处理含噪声数据,具有聚类中心更接近真实聚类中心,聚类准确性高,聚类时间少的优良性能。

A Fast Generalized Noise Clustering （FGNC） based Possibilistic Clustering Algorithm（PCA） is proposed to deal with on Generalized Noise Clustering （GNC） objective function and the shortcoming of GNC algorithm depend heavily on parameters, and FCM must be performed until termination to calculate the parameters for GNC algorithm. With a nonparametric method, FGNC calculates the parameters in GNC objective function. So FGNC algorithm does not depend on the parameters that GNC holds and clusters data faster than GNC algorithm. Experiment and simulation on two man-made data sets and two real data sets show FGNC can deal with noisy data well, cluster centers are closer to real ones, clustering accuracy is improved and clustering time is reduced.