【摘要】 为了解决贝叶斯网络知识合成中的不一致问题,对现有的算法进行了改进.改进后的算法在现有方法的基础上,引入了约束集不一致时的知识合成算法,在知识合成迭代过程中,通过对联合概率分布和概率知识的双向修正,在每一步迭代过程中降低了不一致度,从而实现了贝叶斯网络知识合成时对不一致概率知识的处理.实验结果表明,改进后的算法不仅能够处理概率知识之间相互不一致问题,而且能够处理概率知识与网络结构不一致问题.更多还原

【Abstract】 In order to integrate inconsistent probabilistic knowledge(constraints) into Bayesian network(BN) knowledge integration,an improved algorithm was proposed.In the iterative process of knowledge integration,the new algorithm bi-directionally modifies both joint probability distribution of the BN and probabilistic constraints,thus reducing inconsistency among constraints.Simulation results show that the improved algorithm can handle not only inconsistency among probabilistic constraints but also inconsistency between constraints and the BN structure.更多还原