... P^/a ,) ^ 1, wesee thatI(at; bj) ^ I(at ), (1. 3) and/(a,.; ft, .) ^ I(bj), (1. 4) where J( fl£ ) £ -Iog2 P(a, .), 7(5; .) 4 -Iog2 P(^0,/(a ,) and /(b; -) are defined ... (1.3 0) CH(B/A) & - p(a, b) Iog2 p(b/a)da db, (1.3 1) J — oo J — oc)/*oo TooH(A/B) * - p(a,b)log2p(a/b)dadb, (1.3 2) andH(AB) £ - I I p(a, b)log2p(a, b)dadb, (1.3 3) where ... P(ait bj) = P(ai)P(b j ), then /(«,.; bj) = 0Furthermore, if /(a,•;£>, ) > 0, then P(ai,b j ) > P(ai)P(b j ), there is a higherjoint probability of ai and bj....