摘要
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.通过分析属性相关性的度量和属性约简,选择一组最近似独立的属性约简子集,并结合加权朴素贝叶斯和选择性贝叶斯分类器的优点,提出一种选择性的加权贝叶斯分类器SWNBC.实验结果表明,与朴素贝叶斯分类器相比,WSANBC分类器具有较高的分类正确率.
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. Weighted naive Bayes is it extension. On the basis of analyzing the evaluation of condition attribute with correlation and attribute reduction and selecting a set of more independence attribute, The present paper presents SWNBC（A Selected and Weighted Naive Bayes Classifier） which combined the merits of SNBC and WNBC. Compared with Model, experimental results accuracy. Naive Bayesian Classification show EANBC has higher accuracy.
出处
《湖南文理学院学报（自然科学版）》
CAS
2008年第1期77-79,83,共4页
Journal of Hunan University of Arts and Science(Science and Technology)
基金
安徽省高等学校省级自然科学研究项目(KJ2007B075)
