Are there examples of solving posterior parameter distributions with priors in excel?
I was thinking of logistic regression with 1 parameter and no error term (intercept)that is Y=b*X as an easy example: in excel, logistic regression is solved by using solver to change the values of the parameter [say in cell J1] to maximize the multiplication of the likelihood, =BINOMDIST(Y,1,EXP(linY)/(1+EXP(linY)),FALSE) (1) for all cases, where Y is the binary outcome and linY is the predicted Y from y=b*X. (1) is the likelihood. The prior distribution of the parameter has a normal distribution with a pdf of =NORMDIST(b,mean,stdev,true) (2), where mean and stdev are defined from prior knowledge, lets say mean=0 and stdev=0.05. How do I get my posterior? I know its "proportional" to (1)*(2), prior*likelihood. I have likelihood for each case but the prior is constant. How do I find my posterior - it seems easy question maybe I just haven't thought about it enough.. Thanks! . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
