定性贝叶斯网络中归并计算的进一步探讨及应用

Further Researches and Application of Composition Process in Qualitative Bayesian

在定性贝叶斯网络符号传播算法的基础上,根据实际问题的需要提出了算法中归并运算的改进方法。改进中区分汇连形式各原因节点的逻辑关系,进而提出逻辑或关系在归并过程中应当采用8算子,逻辑与关系采用算子的结论,并通过案例验证了结论在一定范围内的正确性。

Based on the sign-propagation algorithm of qualitative Bayesian networks,an improved composition algorithm is proposed according to actual situation.Two different logic relations of cause-node,logic-or and logic-and,are distinguished.Logic-or relation should use ⊕-operator in the composition process while logic-and should use perator.An example is given to validate the algorithm.