贝叶斯 (Bayesian)网络近年成为数据采掘引人注目的研究方向。通过剖析 Bayesian网络的结构和建造步骤 ,着重讨论用 Bayesian方法从先验信息和样本数据进行学习以确定网络的结构和概率分布的基本方法 ,分析 Bayesian网络学习的特点 ,探讨 Bayesian网络的适用性。与数据采掘的其它方法相比 ,Bayesian网络的优点是可以综合先验信息和样本信息 ,这在样本难得时特别有用 ;可以发现数据之间的因果关系 ,适合于处理不完整数据集 ,这是其它模型难以做到的。其缺点是计算开销较大 ;确定合理的先验密度比较困难 ;如何判定实际问题是否满足所要求的假设 ,没有现成的规则 Bayesian network has become an attractive research direction in data mining in recent years. By analyzing the structure and construction procedure of Bayesian network, the basic method of Bayesian method learning from prior information and sample data to determine the structure and probability distribution of network is analyzed. The characteristics of Bayesian network learning are analyzed, and the applicability of Bayesian network is discussed. Bayesian networks have the advantage of being able to synthesize a priori information and sample information compared to other methods of data mining, which is particularly useful when the sample is rare; causal relationships between the data can be found suitable for handling incomplete data sets, which is Other models are difficult to achieve. The disadvantage is that computational overhead is large; it is difficult to determine a reasonable priori density; how to determine whether the actual problem satisfies the required hypothesis and there is no existing rule