为了认识树形贝叶斯网络(TAN)分类器结构中边的方向与分类精度的关系,以条件对数似然函数为优化目标,研究了TAN分类器结构空间和TAN分类器结构等价类空间,证明了随着训练数据集合的增大,TAN分类器结构中边的方向不对分类器的分类精度产生影响,并提出了一个不考虑边重定向的TAN分类器学习算法,用实验验证了其有效性. In order to understand the relationship between the orientation and the classification accuracy of the TAN classifier structure and the conditional log-likelihood function as the optimization objective, the relationship between the TAN classifier structure space and the TAN classifier structure equivalence class Space, it is proved that with the increase of the training data set, the direction of the side in the TAN classifier structure does not affect the classification accuracy of the classifier, and a TAN classifier learning algorithm that does not consider edge redirection is proposed. Its effectiveness.