一种有效率的关系朴素贝叶斯分类算法

Kind of Efficient Relational Naive Bayesian Classification Algorithm

关系朴素贝叶斯分类算法对于目标关系表和背景关系表中不同的记录关联方式采用不同的策略,灵活运用连接和元组ID传播技术,高效地实现了将背景关系表中的信息加入到目标关系表中一起考虑来进行分类,提高了分类正确率。该算法采用关系数据库的数据表示方式,解决了传统的朴素贝叶斯算法不能支持关系数据库的问题。

The algorithm is relational naive Bayesian classification based on relational model.The algorithm can take the background tables into target table efficiently to improve classification accuracy by treating different tables linked to the target table differently,applying table join and tuple ID propagation techniques.It adopts relational database to represent data,upgrading naive Bayesian classifier to deal with data in relational domain directly.