Regulatory Genes Through Robust-SNR for Binary Classification Within Functional Genomics Experiments

The current study proposes a novel technique for feature selection by inculcating robustness in the conventional Signal to noise Ratio (SNR). proposed method utilizes robust measures of location i.e., “Median” as well variation “Median absolute deviation (MAD) and Interquartile range (IQR)” SNR. By this way, two independent signal-to-noise ratios have been proposed. selects most informative genes/features combining minimum subset genes or features obtained via greedy search approach with top-ranked selected through ratio (RSNR). results are compared well-known gene/feature methods on basis performance metric classification error rate. A total 5 gene expression datasets used study. Different subsets all other included study, their efficacy terms is investigated using classifier models such support vector machine (SVM), Random forest (RF) k-nearest neighbors (k-NN). analysis reveal that (RSNR) produces rates than competing majority cases. For further assessment method, detailed simulation also conducted.