贝叶斯网络结构模型的构建

Construction the Model of Bayesian Network Structure

贝叶斯网络结构是一种将贝叶斯概率方法和有向无环图的网络拓扑结构有机结合的表示模型,它描述了数据项及其依赖关系,并根据各个变量之间概率关系建立图论模型,但是如何获取具有丢失数据的网络结构是一个急需解决的问题.本文提出一个基于Kullback-Leibler(KL)散度的贝叶斯网络结构学习的KLBN(Kullback-Leibler Bayesian Network)算法.实验结果表明,KLBN算法在可靠性方面明显优于传统的具有丢失数据的贝叶斯网络结构学习算法.

A Bayesian network is a graphics model that encodes probabilistic relationships among variables of interest. But it is difficulty to determine the Bayesian network with missing data. In this paper, the KLBN algorithm of learning Bayesian network structure with missing data is presented. Experimental results show that the KLBN algorithm is better than the traditional Bayesian network structure learning algorithm with missing data in reliability.