作者: Peter D. Ho
DOI:
关键词:
摘要: We propose a model-based approach to identifying clusters of objects based on subsets attributes, so that the attributes distinguish cluster from rest population, called an attribute ensemble, may depend being considered. The model is P olya urn model, which equivalent Dirichlet process mixture multivariate normal distributions. This allows for incorporation applicationspecic data features into clustering scheme. For example, in analysis genetic CGH array we account spatial correlation abnormalities along genome.