A REVIEW ON FEATURE SELECTION METHODS AND HIGH DIMENSIONAL GENOMIC DATA

For certain diseases, therapeutic intervention is limited due to varying symptoms, poor medical response and rapid disease development. Genomic studies have revolutionized the way of treating diseases by offering promising ways to detect disease risks, improve preventive methods and offering personalized treatments. The development of laboratory techniques to identify biomarkers generates a vast amount of genomic data. Genomic data comprised different microarray formats and these formats include information pertaining to molecular, biological and cellular levels. The genomic information composed of different micro units which are represented in microarrays. Each microarray is composed of different gene expression features. The choice of features on a microarray data greatly affects the quality of detecting the genes associated with a particular disease. Different methods continuously being investigated to handle high dimensional microarray data and feature selection methods are to distinguish genes from disease and non- disease conditions. This paper aims to review recent feature selection techniques that are applied on microarray dataset. Different dimensionality reduction methods are presented and the performance is analyzed with respect to computational complexity and accuracy in detecting diseases at gene functional level..Research Scholar, Dept. of Computer Science, Karpagam Academy of Higher Education, Coimbatore. Associate Professor, Dept of CS, CA & IT, Karpagam Academy. of Higher Education, Coimbatore.