中文摘要：

分类是数据挖掘领域的重要分支，而贝叶斯分类方法作为分类领域的重要技术得到了日益广泛的研究和应用。限制性贝叶斯网络在不牺牲太多精确性的前提下简化网络结构，是近几年分类领域的研究热点。论文采用统计学中理论较成熟的体积假设检验（Volume Testing）方法寻找属性间的依赖关系，同时结合假设检验的思想和朴素贝叶斯分类算法的优点构造限制性贝叶斯网络，提出了一种基于假设检验的贝叶斯分类算法，并命名为基于体积检验的贝叶斯分类算法。在Weka系统下进行的实验，结果表明，这种方法效果优于朴素贝叶斯方法、TAN算法等，尤其对大数据集有更佳的表现效果。

英文摘要：

Classification is a main branch in Data Mining field.Bayesian classifier as an important technology in this branch has been widely used.Restricted Bayesian learning is a hotspot in these years.In this paper,a kind of hypothesis testing,called volume test is used to find the dependency between attributes.Based on these,propose a method of Bayesian classifier based on hypothe- sis testing,we call it Bayesian classifier based on Volume Test（BVT）.It absorbs advantages of Naive Bayes and idea of statistical hypothesis testing.Experiments show that this method outperforms Naive Bayes,TAN,etc,especially when the dataset is large.