From: oliver@first.in-berlin.de   
      
   Tommy Stromhaug wrote:   
   > Hi,   
   >   
   > Given tha classical alarm, earthquake, burglary, marycalls and johncalls.   
   > Some sort of drawing below.   
   >   
   > My problem is that i want to calculate P(A|M).   
      
      
   Are you shure?   
      
   Or do you want to calc  P(M|A)?   
      
      
   [...]   
   >   
   >        B            E   
   >         \           /   
   >          \         /   
   >            \     /   
   >               A   
   >             /   \   
   >            /     \   
   >          /        \   
   >        J           M   
   >   
   > arrows from B to A and from E to A   
   > arrows from A to J and from A to M   
   >   
      
   Which transitions are known in their likelihood?   
      
   p(B->A) ?   
   p(E->A) ?   
   p(A->J) ?   
   p(A->M) ?   
      
   p(B->J) ?   
   p(B->M) ?   
   p(E->J) ?   
   p(E->M) ?   
      
   And which do you really want to calculate?   
      
   Do you want calc n-steppes likelihoods,   
   or the Bayes term for "how liekely is it that   
   A was connected with M"?   
      
   If the later, then from your diagram   
   you can be sure, that A was the case with p = 1,   
   when you have the state M, because there are no other   
   ways to reach M.   
      
   So I doubt that you want to calc P(A|M).... or   
   maybe your graphic is not complete.   
      
   Ciao,   
       Oliver   
      
   [ 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)