AcSel: selecting variables with accuracy in correlated datasets

Abstract

With the emergence of high-throughput technologies, it is possible to measure large amounts of data relatively at low cost. Such situations arise in many fields from sciences to humanities, and variable selection may be of great help to answer challenges that are specific to each of them. Variable selection may allow to know, among all measured variables, which are of interest and which are not. A lot of methods have been proposed to handle this issue, with the Lasso and other penalized regression as special cases. These methods fail in some cases and linear correlation between explanatory variables is the most common of these, especially in big datasets. In this article, we introduce AcSel, a wrapping algorithm able to enhance the accuracy of any variable selection method. To achieve this result, we use intensive computational simulations.

References in corpus (16)

• Discussion of "Least angle regression" by Efron et al
• Rejoinder to "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Least Angle Regression
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Discussion of "Least angle regression" by Efron et al
• Finding community structure in very large networks
• Nearly unbiased variable selection under minimax concave penalty
• A note on the group lasso and a sparse group lasso
• Bayesian variable selection regression for genome-wide association studies and other large-scale problems
• Random lasso
• A new Universal Resample Stable Bootstrap-based Stopping Criterion in PLS Components Construction