The Graduate Student Seminar Speaker for Tuesday, April 25, 2017 will be Martin Royer.
Title: Clustering methods and the high dimensions
Abstract: This talk focuses on the clustering problem: the task of grouping entities only based on their characteristics according to some measure of similarity. We do an informal survey of classical methods including a focus on the fashionable K-means and its astute reformulation by Peng & Wei (2007). It proves insightful for deriving proofs, understanding other methods as well as the natural difficulties of high-dimensional clustering.