Magnus Fontes
Mathematical Sciences, 
Lund University

Visualization and multivariate analysis

Abstract

We will look at some linear and nonlinear multivariate techniques based 
on singular value decomposition of matrices that form the basis in a 
visualization software developed by the company Qlucore (see www.qlucore.com).
Using the software, we will show how visualization supported by some 
standard statistical stability analysis (for example false discovery rate)
can be very efficient when exploring huge data sets.
Examples will be given from gene expression analysis.