基于特征的贝叶斯过滤网

Feature Based Filtering with Bayesian Networks

以贝叶斯定理为基础，提出并讨论了在特征水平的信息过滤中的贝叶斯过滤网（ＢＮ）的拓扑结构、网节点参数设定、主观概率融合、网节点复合简化、贝叶斯概率推断等问题．研究表明，以ＢＮ构造基于语义特征的信息过滤网。

Based on the Bayesian theorem, we show and discuss the network structuring, parameter settling, subjective probability fusion, compound node making and Bayesian inference in the domain of feature based Bayesian filtering networks. The information filter can help people effectively to retrieve the documents being relevant to a particular information needed from the database.