Microarray-based cancer prediction using single-gene ensemble classifier

Abstract

With the microarray technology widely used in cancer diagnosis, various effective classification approaches have been proposed for gene selection and cancer classification. Generally, single classifiers with numerous genes have been widely used in cancer classification. In biology, many different genotypes can produce the same phenotype. In other words, there might be different pathogenic genes among individuals in various patients. Hence, single classifiers with a plenty of genes suffer from the disadvantage that it is not easy to identify the significant biomarkers for each patient. In this paper, we present a novel approach for cancer classification using ensemble classifier with the single gene to effectively identify candidate pathogenic biomarkers for every cancer sample. We applied our approach to three publicly available cancer datasets and compared classification accuracy to that for six standard methods. The single-gene ensemble classifier performs well in microarray-based cancer prediction.