Margin-based wrapper methods for gene identification using microarray

The gene expression data obtained from microarrays have shown useful in cancer classification problems. DNA microarray data have extremely high dimensionality compared with the small number of available samples. An important step in microarray studies is to select a small number of features (genes) for cancer classifications. In this paper, we propose a novel feature selection algorithm to identify a small set of genes from microarray data for cancer classification. This new wrapper method uses margins to evaluate subsets of genes; bootstrap methods are applied in genetic search to alleviate small sample size problems and support vector machines with maximum margins are used for classifications. The results of experiments carried out show that the proposed method successfully identifies genes informative for cancer classification and provides very reliable cancer classification results.