A Review on Feature Selection Techniques for Gene Expression Data

In disease prediction and diagnosis, the classification of gene expression data plays a vital role. Gene expression data are gained from the microarray technology and it take part a critical role in cancer classification. In gene expression data, every gene does not contribute for efficient classification of gene samples. The developments in DNA microarray technology produces huge number of data samples and the feature dimension of the gene expression data prompt the improvement of a capable and robust algorithm for feature selection in classification of gene expression data. Interpreting the information from the gene expression data is an active area in bioinformatics research and it remains a complicated problem, due to the high dimensional and low sample size. Such problems cause a great challenge to the existing classification techniques. To overcome the challenges in the existing methods, an effective feature selection algorithm is needed to classify the gene expression data. This review article presents an overview of existing feature selection techniques, datasets and performance metrics used for estimating the performance of those algorithms.