Regularized logistic regression with adjusted adaptive elastic net for gene selection in high dimensional cancer classification

摘要: Cancer classification and gene selection in high-dimensional data have been popular research topics genetics molecular biology. Recently, adaptive regularized logistic regression using the elastic net regularization, which is called net, has successfully applied cancer to tackle both estimating coefficients performing simultaneously. The originally used estimates as initial weight, however, this weight may not be preferable for certain reasons: First, estimator biased selecting genes. Second, it does perform well when pairwise correlations between variables are high. Adjusted (AAElastic) proposed address these issues encourage grouping effects real results indicate that AAElastic significantly consistent genes compared other three competitor regularization methods. Additionally, performance of comparable better than Thus, we can conclude a reliable method field classification. showed superior terms all evaluation criteria.The selected more correlated methods.The performed remarkably stability test.In consistency, well.