XPost: comp.ai.neural-nets, sci.image.processing, sci.math.num-analysis   
   From: heath@alumni.brown.edu   
      
   MajorSetback@excite.com wrote:   
   > Phil Sherrod wrote:   
   > > On  2-Dec-2005, MajorSetback@excite.com wrote:   
   > >   
   > > > I would like to separate two classes based upon 8 metrics.  I am   
   > > > thinking of using supervised classification based upon defining a   
   > > > decision hypersurface in 8-dimensional space.  I would be most grateful   
   > > > if someone could suggest the best algorithm for this purpose.   
   > >   
   > > A Support Vector Machine performs hyperplane separation.  A decision tree   
   also   
   > > would be a good method.   
   >   
   > Many thanks for reply.  What I have now is probably like a decision   
   > tree since I have a set of (known) reliability metrics and am seeting   
   > threshold for them based upon a training set.   
      
   That does not generalize well. Set thresholds based on a separate   
   holdout validation set.   
      
   Hope this helps.   
      
   Greg   
      
   > However I think a   
   > decision tree forms hypercubes (if the metrics are orthogonal).  I   
   > would prefer a hypersurface, or perhaps a hyperplane.  However, I will   
   > look into Support Vector Machines.   
   >   
   > Thanks again,   
   > Peter.   
   >   
      
   [ 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)