Using 2D principal component analysis to reduce dimensionality of gene expression profiles for tumor classification

Abstract

In the last ten years, numerous methods have been proposed for accurate classification of tumor subtype based on gene expression profiles (GEP). Among these methods, feature extraction methods play an important role in constructing classification model. However, traditional methods view a gene expression sample as 1D vector, which does not sufficiently utilize the correlation and structure information among many genes. We, therefore, introduce 2D principal component analysis (2DPCA) to extract features for tumor classification by converting 1D sample vector into 2D sample matrix. To evaluate its performance, we perform a series of experiments on four tumor datasets. The experimental results indicate that the obtained performance by using 2DPCA is superior to the classic principal component analysis. © 2012 Springer-Verlag.