基于树选择的TAN分类器

通过对朴素贝叶斯（NBC）分类器与传统的基于树扩展的贝叶斯（TAN）分类器的分析，对TAN分类器进行改进，提出CTAN分类器。朴素贝叶斯分类器对非类属性独立性进行完全独立假设，传统TAN则弱化所有属性的独立性．提出的CTAN则是通过操作TAN保留对数对部分相关属性有选择的进行弱化。CTAN改进的方向主要是对属性关系树的部分利用，通过实验证明，CTAN要优于传统TAN分类器。

Tree Augmented Naive Bayes Based on Choice-Tree

Improves the TAN and puts forward CTAN through the analysis of Naive Bayesian Classifier and Tree Augmented Naive Bayes.Naive Bayes classifiers assumptions about the attributes being completely independent. Tree Augmented Naive Bayes weakens independence of all the atrributes. CTAN weakens independence of partial attributes by operating the reserved number. The improvement direction of CTAN is the utilization of relational tree. Through experiments prove, CTAN is superior to TAN.