Swarm intelligence based feature Selection methods and classifiers For cervical cancer classification in Gene expression

Abstract

In the encompassing world, cervical squamous cell carcinoma is foresighted to be ubiquitous cancer having fourth rank and also obligatory death among women. According to World Health Organization (WHO), 86% of cervical cancer cases are reported across developing countries. Meanwhile, prostrate income countries account for the highest monotonous antecedent pertaining to cancer mortality. This cervical carcinoma springs gainst the cervix owing to the aberrant proliferation of cells with the potentiality to plunder or pervade supplementary organs of the body. Gene expression profiling is still extensively employed in cervical cancer research. Gene expression patterns have been explored in biological networks, especially a gene co-expression network that crops up as an offbeat comprehensive technique for microarray investigation. In gene co-expression networks, a part of physiognomy remains disproportionate at the moment of peculiar biological complication is resolved. Feature selection is considered a crucial aspect in identifying a portion of genes that yields better results based on this subset of features. Traditional gene selection techniques could not cultivate the finest potential series of genes which might reduce the accuracy of the classifier. For gene selection, numerous researchers have promoted optimization techniques. The new feature selection algorithm complies based on the metaheuristic algorithm and it is appropriate for choosing a subset of genes. Traditional Differential Co-expression Networks (DCNs) speculate that the underlying gene expression related to Differentially Expressed Genes (DEGs) and normal data are not evenly distributed. However, the optimal selection of genes or recurrence genes in the prediction also becomes a very crucial task newline