Hi Peter (and list) For the variable selection problem in large-p-small-n problems we are faced with (e.g., microarrays) we have been developing (very slowly) an approach using genetic algorithms and Mantel correlation. I have attached a proceedings paper to appear this Sept. in the Interface proceedings and John Grefenestette's powerpoint file from the 2005 CSNA/Interface meeting. Peter, this might be an approach for your 900 variables. Bill Peter Flom <[log in to unmask]> wrote: William Shannon wrote (Peter, you may not be interested in covariate selection since your data will always consist of the 900 measurements.) I actually am interested in variable selection. One of our tasks here is to find a small subset of the 900 measurements that works reasonably well. Peter ---------------------------------------------- CLASS-L list. Instructions: http://www.classification-society.org/csna/lists.html#class-l ---------------------------------------------- CLASS-L list. Instructions: http://www.classification-society.org/csna/lists.html#class-l