利用基因表达数据提出一种新的网络模型—贝叶斯网络,发现基因的互作.一个贝叶斯网络是多变量联合概率分布的有向图模型,表示变量间的条件独立属性.首先我们阐明贝叶斯网络如何表示基因间的互作,然后介绍从基因芯片数据学习贝叶斯网络的方法. Using the gene expression data, we propose a new network model, Bayesian Networks, to discover the interaction of genes. A Bayesian network is a directed graph model of multivariate joint probability distribution, which indicates the conditional independence between variables. First, we Clarify how Bayesian networks represent inter-gene interactions and then introduce methods to learn Bayesian networks from gene chip data.