67第五章不确定性推理方法(简s2)

•••Bayes•••Bayes••••••••SCELBADEFCGBADCEGF••––•••mutuallyexclusiveandcollectivelyexhaustivestates•••“”••••XYP(X)P(Y|X)P(X,Y)=P(X)P(Y|X)XZYP(X)P(Z|Y,X)P(Y)P(X,Y,Z)=P(X)P(Y)P(Z|X,Y)••••¬•¬•¬¬SCELP(S)=0.4P(C)=0.3P(E|S,C)=0.9••••••••DSCEL)(*)|(*),|(*),,|(),,,(SPSCPCSLPLCSEPELCSP=)(*)(*)|(*),|(),,,(SPCPSLPCSEPELCSP=D•••••SCELP(S)=0.4P(C)=0.3P(E|S,C)=0.9D-•••XZYD-••…••……NYXZD()•••……ZNYXD()ZNYXeZNYXLMeD()•Bayes–––()D()•)|(),|(εεijiVPVVP=)|(),|(εεjijVPVVP=)|,(εjiVVI)|,(εijVVID()•–•–•–D(1)XYXYXYYesYesZXYXYXYYesNoZXYXYXYYesNoZXYXYXYNoYesZZXYXYXYNoNoD(2)ZXYX—Y—Z—P(X,Y)P(X)P(Y)P(X|Y,Z)=P(X|Z)ZXYX—Y—Z—P(X,Y=P(X)P(Y)P(X|Y,Z))P(X|Z)D(3)X—Y—Z—W—P(X,Y)=P(X)P(Y)P(X|Y,Z)=P(X|Z)YZWXX—Y—Z—W—P(X,Y)P(X)P(Y)P(X|Y,Z)P(X|Z)YZWXD(4BatteryRadioIgnitionGasMovesStart•–•––••–––••––––•–––•P(X)P(Y)Query:P(X|Y)P(X)P(Z|X,Y)Query:P(X|Y,Z)andP(X|Z)P(Y)•–––•–P(Y|X)Query:P(X|Z)P(X)P(Z|Y)P(E|S)SEECP(E|S)=P(E,C|S)+P(E,~C|S);P(E,C|S)=P(E,C,S)/P(S)P(E|C,S)*P(C,S)/P(S)=P(E|C,S)*P(C|SP(E,~C|S)=P(E|~C,S)*P(~C)P(E|S)=P(E|C,S)*P(C)+P(E|~C,S)*P(~C)=0.9*0.3+0.3*(1-0.3)=0.48SCELSCELP(~C|~E),P(~C|~E)=P(~E|~C)*P(~C)/P(~E)()P(~E|~C)=P(~E,S|~C)+P(~E,~S|~C)=P(~E|S,~C)*P(S)+P(~E|~S,~C)*P(~S)=0.82P(~C|~E)=0.82*(1-0.3)/P(~E)=0.574/P(~E)P(C|~E)=0.102/P(~E)P(~C|~E)+P(C|~E)=1,P(~E)=0.676P(~C|~E)=0.849SCELP(~C|~E,~S),=0.807