ABSTRACT
This chapter provides an overview of the application of the various feature selection and feature extraction techniques commonly used in bioinformatics. The key areas describe the issues and challenges faced during the analysis of high–dimensional data, whether gene expression data, protein sequence, or structural data. Nearly all normalization techniques are bases on the assumption that one or more control genes are constitutively expressed at near–constant levels under all experimental conditions. Local normalization strategies take into consideration subsets of array elements; normalization is then performed on independent subsets. The primary applications of microarray technology are to analyze genes from different samples and identify differentially expressed genes between samples. In gene expression analysis, face the problem of constructing an accurate prediction rule R using a dataset consisting of a relatively small number of microarray samples, with each sample containing the expression data of many genes. To resolve the issue of semantics in bioinformatics databases, ontologies have provided better biological interoperability.
