文件名称：Attribute Weighting for Averaged One-Dependence Estimators
文件大小：373KB
文件格式：PDF
更新时间：2022-03-15 02:54:03
Bayesian Est naive bayes attribute
这是我写的一个论文，关于朴素贝叶斯分类器经过属性权重调整提升分类性能并提出了新算法，名字是WAODE，摘要如下： Averaged One-Dependence Estimators (AODE) is one of supervised learning algorithm, which relaxes the conditional independent assumption that working on the standard naive Bayes learner. The AODE has shown the reasonable improvement of classified property compared to naive Bayes learner. However, AODE does not consider the relations between super-parent attribute with other normal attributes. In this paper, we propose a method based on AODE which weighted the relationship between the attributes. We proposed a weighted AODE (WAODE), which was an attribute weighting method which employs conditional mutual information metric to rank the relations among the attributes. We have conducted experiments on UCI benchmark data sets and made a comparison of accuracies between AODE and our proposed learners. The experimental results in our paper have shown that WAODE had better performance in accuracy than original AODE.