Quote:
Originally Posted by squeek.
Yeah, that's the next step. I thought it wouldn't matter since all I care about is the relative values of the Bayesian average (all due respect to the Bayesian people and their culture), not the Bayesian average itself (on Wikipedia it says: "In cases where the averages' relative values are the only result of importance, m can be replaced with zero."), but I'll fix the formula and see what happens.
|
Hmm, relatively speaking, I don't think it'll matter then. Simple analysis would dictate that each average would just have a constant factor added, which means the order is the same. That is, for A/f > B/f > C/f implies A/f+c/f > B/f+c/f > C/f+c/f. By the way, I don't understand how you can get 100%. It shouldn't even be possible if you're using bayes...