Topic 1 – Probability Module 1.2 – Probability Theory Total Probability Rule & Bayes Rule Suppose we have events X1, X2, …, Xk where Xi∩Xj = ∅ for i ≠ j, and an event Y where Y ⊆ X1∪X2∪…∪Xk where ⊆ denotes subset. (As an example, visualize the X’s as the states of the US and Y as a region of the US, e.g., the South West). TOTAL PROBABILITY RULE: If Y ⊆ X1∪X2∪…∪XK and Xi∩Xj = ∅ for i ≠ j, then P(Y) = P(Y∩X1) + P(Y∩X2) + … + P(Y∩XK). By the general multiplication rule, applied k times, we have If Y ⊆ X1∪X2∪…∪XK and Xi∩Xj = ∅ for i ≠ j, then P(Y) = P(Y/X1)P(X1) + P(Y/X2)P(X2) + … + P(Y/XK)P(XK). BAYES RULE (SIMPLE FORM): If Y ⊆ X1∪X2∪…∪XK and Xi∩Xj = ∅ for i ≠ j, P(X i )P(Y /X i ) then P(X i /Y) = P(Y) By the total probability rule, we have: € BAYES RULE (EXTENDED FORM): If Y ⊆ X1∪X2∪…∪XK and Xi∩Xj = ∅ for i ≠ j, P(X i )P(Y /X i ) then P(X i /Y) = K ∑ P(X j )P(Y /X j ) j=1 € Bayes rule is the rule by which we revise probabilities; it will play a crucial role in decision analysis and the economic theory of information. For example, suppose we have P(X1) = .2 P(X2) = .3 P(X3) = .5 Xi∩Xj = ∅ for i ≠ j; i, j = 1, 2, 3 Y ⊆ X1∪X2∪X3 P(Y/X1) = .4 P(Y/X2) = .6 P(Y/X3) = .8 Then, by the total probability rule, we have P(Y) = P(Y/X1)P(X1) + P(Y/X2)P(X2) + P(Y/X3)P(X3) = (.4)(.2)+(.6)(.3)+(.8)(.5) = .66. € By Bayes rule, we have P(X1 )P(Y /X1 ) (.2)(.4) .08 P(X1 /Y) = = = P(Y) .66 .66 P(X 2 /Y) = P(X 2 )P(Y /X 2 ) (.3)(.6) .18 = = P(Y) .66 .66 P(X 3 /Y) = P(X 3 )P(Y /X 3 ) (.5)(.8) .40 = = P(Y) .66 .66 € € Note that the denominator, .66, is simply the sum of the numerators, .08, .18, and .40. € PROBLEM: Suppose we have P(X1) = .2 P(X2) = .3 P(X3) = .5 Xi∩Xj = ∅ for i ≠ j; i, j = 1, 2, 3 Y ⊆ X1∪X2∪X3 P(Y/X1) = .1 P(Y/X2) = .5 P(Y/X3) = .9 Find P(Y), P(X1/Y), P(X2/Y), and P(X3/Y)
© Copyright 2026 Paperzz