Dear All, We have recently published an article on cross-validation pitfalls http://www.jcheminf.com/content/6/1/10/ In the article, amongst other things, we discuss nested cross-validation and variable selection with cross-validation which I am sure is of interest to statistical modelling community. I am keen to hear critical assessment of our paper from the community. So far it was brought to my attention thata) In the section "Preprocessing" I failed to mention that we used the caret R package and failed to reference Max Kuhn's 2008 paper "Building Predictive Models in R Using the caret Package".b) In the section "Nested cross-validation" I failed to mention the Bioinformtics 2004 paper by Statnitkov et al. "A comprehensive evaluation of multicategory classification methods...", because they also suggested nested cross-validationc) In the section "Discussion" I should have explained that nested cross-validation may also be applied to cases where variable selection is performed with cross-validation. There is a Bioinformatics 2010 paper by Lagani and Tsamardions "Structure-Based Variable Selection for Survival Data" where they applied nested cross-validation for variable selection with cross-validation. Looking forward to hearing from you.DKP.S. I am the corresponding author and please use my email address specified in the paper. [[alternative HTML version deleted]]
______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
