Abstract
This talk discusses and demonstrates interactive visual methods on the data from a paper by Dudoit, Fridlyand and Speed (2002). The data was collected on lymphona, leukemia and of the National Cancer Institute's 60 cell lines. The paper compares the performance of classification techniques for microarray data on tumors.

This data poses a problem for statistical analysis and visualization because there are considerably more variables (genes) than cases (microarray experiments). Nevertheless there are some details to learn about the differences of the tumor arrays based on gene expression values. The graphics helps us to understand how the classification methods perform on the data.