From: jm@alb.fr   
      
   Dear all,   
      
   I am a student and I have to do a strong Othello software.   
   I studied Michael Buro's reports, I implemented his technics using   
   patterns and logistic regression but I have still a 25-30 % of   
   prediction error, I think is too much.   
      
   Here are the vectors of my regression :   
      
   pattern1_value, pattern2_value,  ... , pattern52_value,  class   
      
   where   
   patternx_value is the Y+0.5/(N+1.0) - 0.5,  with Y=the number of   
   victories for this pattern, N the total number of positions where this   
   pattern appear   
      
   class = 1 if this position has a majority of victories, 0 otherwise   
      
   I use 13 differents pattern, and with 4 symmetries, I have a 52 vector   
   values.   
      
   Is it the good way to proceed, or I made something wrong ?   
      
   Thanks for your answer...   
      
   JM   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)