From: mailbox@dmitry-kazakov.de   
      
   On 18 Oct 2004 11:32:00 -0700, Ethan Seng wrote:   
      
   > In regard to my previous post, the inputs C's and V's for each of the   
   > M datasets can be obtained from fitting the equation   
   > 'ln(y)=ln(30/V)-(C*t/V)' to the k data pairs.   
      
   Excellent, it is linear now.   
      
   > These data (to be   
   > collated and combined (in some way) from the M sets of data) together   
   > with the y's and t's then form the input base from which neurofuzzy   
   > algorithms are applied.   
      
   Because it is linear you can just use standard linear regression. It should   
   be definitely better than any other least squares approximation.   
      
   But the actual problem as far as I understand you is that approximation is   
   only the first step. After that you should blur the approximation so that   
   it will fit all data exactly. Then you will have V and C fuzzy.   
      
   --   
   Regards,   
   Dmitry A. Kazakov   
   http://www.dmitry-kazakov.de   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)