Kernel clustering methods have been used successfully to cluster non linearly separable data. In this paper, we propose a modification of the Kernel K-means, called the Multi-Scale Kernel K-means, that addresses one important challenge, which is the automated estimation of the kernel scale parameters for data containing clusters with different scale values. We propose a novel method that estimates the local kernel scales using the local data density in the original space to learn an adaptive and localized kernel function. Our experimental results with the Multi-Scale Kernel K-means show significant enhancements over the standard Kernel Kmeans for data sets containing clusters with varying scales and densities.
展开▼