B0886
Title: Variable selection for clustering
Authors: Jeffrey Andrews - University of British Columbia (Canada) [presenting]
Abstract: Variable selection under a clustering paradigm is discussed, highlighting recent advances in the literature and available software. The VSCC (variable selection for clustering and classification) method looks for variables which minimize the within group variation, while simultaneously ensuring that redundant features are removed. The technique and R software are introduced, focusing on some recent updates, and then comparisons with state of the art methods on a variety of data sets, both real and simulated, are shown.
