基于TAN结构的贝叶斯文本分类器研究

摘要

Nave Bayesian is a simple and effective method that can be easily implemented text automatic categorization,but its assumption of independence does not exist in practice.In TAN structure Bayesian classification algorithm,the correlation between each two attributes was considered,the assumption of independence between attributes was reduced in a certain degree.%朴素贝叶斯分类器是一种简单且有效实现的文本自动类方法,但其独立性假设在实际中是不存在的。在TAN结构贝叶斯分类算法中,考虑了两两属性间的关联性,对属性间的独立性假设有了一定程度的降低。