XPost: comp.ai.neural-nets, sci.math.num-analysis, sci.math   
   From: ted.dunning@gmail.com   
      
   Greg Heath wrote:   
   > > Ted Dunning wrote:   
   > > It doesn't really solve the problem,   
      
   > I assume you are referring to *Linear* PCA and PLS.   
      
   Actually, no, I am referring to the fact that finding a low dimensional   
   sub-space in which you can build a classifier is not really the answer   
   to the problem of building a high-dimensional classifier.   
      
   I agree that it often works, but whether or not dimensionality   
   reduction works depends on whether or not there is a usable   
   low-dimensional sub-space at all.  This is by no means guaranteed even   
   if it often works.   
      
   [ comp.ai is moderated.  To submit, just post and be patient, or if ]   
   [ that fails mail your article to , and ]   
   [ ask your news administrator to fix the problems with your system. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)