The HPBNET Procedure

Probability Table

The following statements produce Figure 4.2, which shows the conditional probability table for each node in the network. You can use these probability tables for scoring or inferences or both. The conditional probability tables together with the network structure determine the Bayesian network.

 proc print data=network noobs label;
     var _parentnode_ _parentcond_ _childnode_ _childcond_ _value_;
     where _type_="PROBABILITY";
 run;

Figure 4.2: Probability Table

Parent Node	Parent Condition	Child Node	Child Condition	Value
		Treatment	A	0.33333
		Treatment	B	0.33333
		Treatment	P	0.33333
Treatment	A	Pain	YES	0.27273
Treatment	A	Pain	NO	0.72727
Treatment	B	Pain	YES	0.27273
Treatment	B	Pain	NO	0.72727
Treatment	P	Pain	YES	0.72727
Treatment	P	Pain	NO	0.27273
Pain	YES	Age	<67	0.17857
Pain	YES	Age	<75	0.35714
Pain	YES	Age	>=75	0.46429
Pain	NO	Age	<67	0.31579
Pain	NO	Age	<75	0.60526
Pain	NO	Age	>=75	0.07895