I.: Normalized compression distance for gene expression analysis

Abstract — In this paper we show that the normalized compression distance can be applied to gene expression data analysis. Typically, microarray-based classification involves using a feature subset selection method in connection with a specific distance metric. The performance is dependent on the selection of the methods. With our proposed approach there is no need for feature subset or distance metric selection and all the data can be used directly with the universal similarity metric. We demonstrate the method on simulated and real microarray data. I