I have a problem which I think (not compeletly sure) could be managed by soft evidences.suppose we have three nodes (A,B,C).A and B are parents of C.

                    A   B       
                      \ /
                      C
suppose we have 4 states (S1,S2;S3,S4) and
P(S1 |A)=.2  P(S2 |A)=.15   P(S3|A)=.15  P(S4|A)=.5
P(S1 |B)=.3  P(S2 |B)=.4    P(S3|B)=.1   P(S4|B)=.2
what are P(S1 |C)  P(S2 |C)   P(S3|C)  P(S4|C) ?

(Posterior probability of C knowing other two observations)
Please let me know how I can solve the problem using Kevin Murphy Toolbox?