Abstract: MicroRNAs are small non-coding RNA molecules which are important developments in the cancer biology. miRNA microarrays are useful tools to identify potential biomarkers for variety of cancers. Due to high dimensionality of microarrays, it is very hard to identify cancer oncogenes and classify tumor samples. Feature selection is very essential task in the process of classification and identification of biomarker genes by selecting relevant genes. In this research, Entropy Based Mean Score (EBMS) is employed to identify the biomarker genes in miRNA microarrays. This is based on Fisher score which has the benefits of information gain and achieves maximum classification accuracy. The proposed research is tested on benchmark datasets with SVM and ANN for classification. The experimental results show that the EBMS method outperforms the existing methods and it is suitable for effective feature selection.
M. Anidha and K. Premalatha, 2017. An Entropy based Mean Score Feature Selection Method for Identification of Biomarkers using Mirna Expression Profiles for Cancer Classification. Asian Journal of Information Technology, 16: 206-211.