Inference Using Full Joint Distributions
Consider three Boolean random variables A, B, and C. We abbreviate the propositions A = true, B = true, C = true as a, b, c, respectively, and A = false, B = false, C = false as ¬a, ¬b, ¬c, respectively. Given below is a table showing their joint probability distributions.
(a) Compute the conditional probability distribution P(A | ¬c) = ⟨P(a | ¬c), P(¬a | ¬c)⟩. (To normalize a vector, say, ⟨x, y⟩, you may write α⟨x, y⟩, which represents ⟨x/(x+y), y/(x+y)⟩.)
(b) Compute the conditional probability distribution P(B | ¬c) = ⟨P(b | ¬c), P(¬b | ¬c)⟩.
(c) Are A and B conditionally independent given C? Explain why or why not.