一种基于强属性限定的加权贝叶斯分类器

朴素贝叶斯分类器是一种简单而高效的分类器,但它的条件独立性假设使其无法将属性间的依赖关系表达出来,影响了它分类的正确率,加权朴素贝叶斯是对它的一种扩展.基于强属性限定的贝叶斯分类器SANBC,通过在强弱属性之间添加增强弧以弱化朴素贝叶斯的独立性假设,扩展了朴素贝叶斯分类器的结构;结合加权朴素贝叶斯和基于强属性限定的贝叶斯分类器SANBC的优点,提出一种基于强属性限定的加权贝叶斯分类器WSANBC;实验结果表明,WSANBC分类器具有较高的分类正确率.

A weighted and restricted Bayesian classifier based on strong attributes

The naive Bayesian classifier is a simple and effective classifier,but its attribute independence assumption makes it unable to express the dependence among attributes and affects its classification accuracy.The weighted naive Bayes is an extension of it.The SANBC which is a restricted Bayesian classifier based on strong attributes extends the structure of the naive Bayesian classifier through the adding of highlighting lines between strong and weak attributes so that the naive Bayesian classifier can be weakened.The present paper presents the WSANBC which is a weighted and restricted Bayesian classifier based on strong attributes and combines the merits of SANBC and WNBC.Experimental results show that the WSANBC has higher accuracy.