AI5章-不确定性推理.ppt

解：由已知知识得到的推理网络如下图所示。 6.5.1 DS理论的形式描述 1. 概率分配函数（1/3） (1) 求CER(A) ∵ CER(E1 AND E2) =min{CER(E1), CER(E2)} =min{0.8, 0.6} =0.6 m({a1}, {a2})={0.6×0.3, 0.6×0.5} ={0.18, 0.3} Bel(A)=m({a1})+m({a2})=0.18+0.3=0.48 Pl(A)=1-Bel(﹁A)=1-0=1 f(A)=Bel(A)+|A|/|Ω|*[Pl(A)-Bel(A)] =0.48+2/10*[1-0.48] =0.584 ∴ CER(A)=MD(A/E)×f(A)=0.584 (2) 求CER(B) ∵ CER(E3 AND (E4 OR E5)) =min{CER(E3), max{CER(E4), CER(E5)}} =min{0.9, max{0.5, 0.7}} =min{0.9, 0.7}=0.7 m({b1})=0.7×0.7=0.49 Bel(B)=m({b1})=0.49 Pl(B)=1-Bel(﹁B)=1-0=1