视觉采样聚类方法VSC

基于视觉采样原理，该文提出了一般化的视觉采样聚类方法VSC。该方法将视觉原理与著名的Weber定律结合起来，其特点是:对聚类初始条件不敏感；Weber定律提供了新的聚类有效性标准，并且该方法所得到的合理的聚类数可以依据Weber定律而得到。大量的实验结果表明了算法VSC的有效性。文中讨论了算法VSC与由Yang Miin-Shen等人(2004)新近提出的基于相似度量的聚类算法SCA之间的内在联系，得出了这两个算法具有一定的同解性质，从而揭示了该文所提方法VSC能够有效地克服算法SCA中参数 不易确定的困难。

Visual Sampling Based Clustering Approach VSC

Based on the visual sampling principle, the generalized visual sampling based clustering approach VSC is proposed. The clustering approach incorporates the visual sampling principle with the famous Weber law such that it has two distinctive advantages: firstly, it is insensitive to initial conditions; secondly, the reasonable clustering number can be effectively determined by the new Weber-law-based clustering validity index. The experimental results demonstrate its success. Moreover, the link relationship between our approach and algorithm SCA (Similarity-based Clustering Algorithm) recently proposed by Yang Miin-Shen, et al. (2004) is derived. Both theoretic analyses and experimental results show that in many cases, the approach here has almost the same clustering results as algorithm SCA. This fact reveals that the approach can be used to overcome the drawback of SCA, i.e., the parameter y is very difficult to be well determined.