From: walter@bytecraft.com   
      
   David Tian wrote:   
   >   
   > Hi,   
   >   
   > I have got a problem at hand. It is a sales forecasting problem. In   
   > the data set both input and output variables are continuous valued.   
   > The input variables describe the characteristics of a supermarket   
   > store for example. They could be the store location, store size, range   
   > of items held, and so on. The output variable is an indicator of sales   
   > for example, number of £'s per month. Say the sales figure is between   
   > £0.00 and £1000.00. Say I group the values of output variable into   
   > different intervals, say, 'low', 'medium' and 'high'. Then is it   
   > possible to induce classification rules from the continuous data in   
   > the following way using a regression tree.   
   >   
   > Regression Data --> Regression Tree --> Regression Rules -->   
   > Discretization --> Classification Rules   
   >   
   > The discretization step groups the values of output variable into the   
   > intervals, 'low', 'medium' and 'high'.   
   >   
   > Are there any standard/popular approaches for this type of problems?   
   > Are fuzzy regression trees suitable for this type of problems? If   
   > there are standard approaches, may I have some suggestions of the   
   > relevant/key papers/references in this area?   
   >   
   > Regards,   
   >   
   > David   
   >   
      
   There is a fuzzy logic example on Byte Craft's website that has a   
   pricing example that might be useful in understanding a fuzzy logic   
   approach.   
   http://www.bytecraft.com/fuzzy.html   
      
   If you have any questions email me off line.   
      
   w..   
      
   [ 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)