Support Vector Machines and Neural Networks as Marker Selectors for Cancer Gene Analysis

DNA micro-array analysis allows us to study the expression level of thousands of genes simultaneously on a single experiment. The problem of marker selection has been extensively studied but in this paper we also consider the quality of the selected markers. Thus, we address the problem of selecting a small subset of genes that would be adequate enough to discriminate between the two classes of interest in classification, while preserving self-similar characteristics to allow closed clustering within each class.