Cluster-based visualisation with scatter matrices

Authors:
P. J. G. Lisboa;I. O. Ellis;A. R. Green;F. Ambrogi;M. B. Dias
Affiliations:
School of Computing and Mathematical Sciences, John Moores University, Liverpool L3 3AF, UK;Department of Histopathology, School of Molecular Medical Sciences, Nottingham University Hospitals Trust, University of Nottingham, NG5 1PB, UK;Department of Histopathology, School of Molecular Medical Sciences, Nottingham University Hospitals Trust, University of Nottingham, NG5 1PB, UK;Department of Preventive and Predictive Medicine, Istituto Nazionale per lo Studio e la Cura dei Tumori, 20133 Milano, Italy;Mathematics and Psychological Sciences, Unilever Corporate Research, Sharnbrook, Bedfordshire MK44 1LQ, UK
Venue:
Pattern Recognition Letters
Year:
2008

Citing 3
Cited 4

GTM: the generative topographic mapping

Neural Computation
A Nonlinear Mapping for Data Structure Analysis

IEEE Transactions on Computers
Selective smoothing of the generative topographic mapping

IEEE Transactions on Neural Networks

Feature and model selection with discriminatory visualization for diagnostic classification of brain tumors

Neurocomputing
Application notes: data mining in cancer research

IEEE Computational Intelligence Magazine
Cohort-based kernel visualisation with scatter matrices

Pattern Recognition
Identifying malignant transformations in recurrent low grade gliomas using high resolution magic angle spinning spectroscopy

Artificial Intelligence in Medicine

Quantified Score

Hi-index	0.10

Visualization

Abstract

The trace of the scatter matrix, which measures separation between population cohorts, is shown to be strictly preserved by sphering the data followed by a projection onto the space of population means. This result suggests using the space of means as a basis to calculate well-separating lower-dimensional projections of the data, derived from the scatter matrix in the projective space. In particular, it defines an approximation to the canonical decomposition of the scatter matrix that applies for singular covariance matrices. The method is illustrated with reference to k-means clusters in data sets from bioinformatics and marketing.