That is I have likelihood P(D|H) for each case and the prior is a known distribution. I guess it must involve sampling the prior distribution for several values *for each case*, say 10 =NORMDIST() columns and averaging them out? which sounds very computer intense. If this is the case, maximum likelihood estimation is not even needed?
A non-technical strategy for sampling/solving this problem appreciated, I'm sure by other newbies too! . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
