Boley, DanielBorst, Vivian2020-09-022020-09-021999-07-23https://hdl.handle.net/11299/215385Fast and effective unsupervised clustering is a fundamental tool in unsupervised learning. Here is a new method to explore large datasets that enjoys many favorable properties. It is fast and effective, and produces a hierarchical structure on the underlying dataset, without using a training set. It also yields auxiliary information on the significance of the different attributes.en-USReport