Introduction Of Bayesian Network课件.pptVIP

Introduction Of Bayesian Network Lab202 , 2006/10/04 精品文档 Bayesian Network Bayes Rule P(B|A) = P(A,B) P(A) P(A,B)=P(B|A)*P(A)=P(A|B)*P(B) P(S1|E)=P(E|S1)*P(S1) P(E) 精品文档 Bayesian Network Diagram of Bayes Theorem 精品文档 Bayesian Network Regarded as the node of the network and made up by a group of random parameters, can state the probability relation of one of parameters, often use in the analysis question 精品文档 Bayesian Network and Probability Form Cloudy Sprinkler Rain Wet Grass C P(S =T) P(S =F) T 0.2 0.8 F 0.3 0.7 P(C =T) P(C =F) 0.1 0.9 C P(R =T) P(R =F) T 0.4 0.6 F 0.6 0.4 SR P(W =T) P(W =F) TT 0.1 0.9 TF 0.2 0.8 FT 0.4 0.6 FF 0.6 0.4 精品文档 Bayesian Network Yes 0.1 No 0.9 Yes 0.58 No 0.42 Yes 0.29 No 0.71 Cloudy Sprinkler Rain Wet Grass Yes 0.385 No 0.615 精品文档 Bayesian Network Cloudy Sprinkler Sprinkler 0.1 T 0.9 F 0.8 F 0.2 T 0.3 T 0.7 F P(C,S) =0.02 P(C,! S)=0.08 P(! C,S) =0.27 P(! C,! S)=0.63 P(S) =0.02+0.27 =0.29 P(! S)=0.08+0.63 =0.71 精品文档 Bayesian Network Cloudy 0.1 T 0.9 F 0.6 F 0.4 T 0.6 T 0.4 F P(C,R) =0.04 P(C,! R)=0.06 P(! C,R) =0.54 P(! C,! R)=0.36 P(R) =0.04+0.54 =0.58 P(! R)=0.06+0.36 =0.42 Rain Rain 精品文档 Bayesian Network Sprinkler 0.58 T 0.42 F 0.71 F 0.29 T 0.1 T 0.9 F P(S,R,W) =0.1682 P(W) =0.38482 P(! W)=0.61518 Rain Rain Wet Grass Wet Grass Wet Grass Wet Grass 0.58 T 0.42 F 0.2 T 0.8 F 0.6 T 0.4 F 0.4 T 0.6 F P(S,R,!W)=0.15138 P(S,!R,W) =0.02436 P(S,!R,!W)=0.09744 P(!S,R,W) =0.16472 P(!S,R,!W)=0.24708 P(!S,!R,W) =0.17892 P(!S,!R,!W)=0.11928 精品文档 Properties Ones that concerned had all already to let had directionality that is linked or arrow number contact Every node have one terms probability form, in order to prove nodal father node this to this nodal influence Do not allow to have directionality circulation to exist in the network Need t