Towards a Memetic Feature Selection Paradigm [Application Notes]

Feature selection has become the focus of many real-world application oriented developments and applied research in recent years. With the rapid advancement of computer and database technologies, problems "with hundreds and thousands of variables or features are now ubiquitous in pattern recognition, data mining, and machine learning [1], [2]. In this article, we consider two real-world feature selection applications: gene selection in cancer classification based on microarray data and band selection for pixel classification using hyperspectral imagery data.