Re: Dealer Rating Analysis

From Stanford Wong's BJ21

Posted by Pete Moss on 30 Dec 1997, at 2:33 a.m., in response to Dealer Rating Analysis, posted by $mart Gambler on 28 Dec 1997, at 11:01 p.m.

There's another bit of information that's needed. Say you want to figure the probability P(A|WHG), the probabilty of getting an A dealer, given the dealer is a white-haired granny (WHG). The Bayes theorem, [Revd. Thomas Bayes, 1702-1761], says, 


    P(A|WHG) = P(WHG|A) * P(A)/P(WHG)

P(A) is the frequency of A's amoung all dealers that you consider, which is 16.3% estimated from your data. P(WHG) is the frequency of of white-haired grannies among all the dealers you consider. Gotta know that.

Pete

Responses