B0592
Title: Clustering three-way data
Authors: Paul McNicholas - McMaster University (Canada) [presenting]
Abstract: Some approaches to clustering three-way data are presented. These approaches, which are based on mixtures of matrix-variate distributions, include a method for high-dimensional data and another for dealing with outliers. Selected approaches are illustrated on simulated and real data. Conclusions with some discussion about directions for future work are drawn.
