一种扩展的朴素贝叶斯分类器改进算法  

An Improved Algorithm for Learning Augmented Naive Bayes Classifier

作　　者：张丽霞[1] 赵大宇[1] 

机构地区：[1]沈阳师范大学数学与系统科学学院,辽宁沈阳110034

出　　处：《计算机技术与发展》2006年第5期28-30,共3页Computer Technology and Development

基　　金：辽宁省高等学校科学研究项目(202112020)

摘　　要：文中研究贝叶斯分类器家族中的一种扩展朴素贝叶斯分类器。此种扩展朴素贝叶斯分类器满足两个条件:一是类结点是所有属性的父结点;二是每个属性最多有一个属性父结点。其中有代表性的两种算法是贪婪爬山算法(Hill Climb-ing Search,即HCS算法)和超父结点算法(Superparent,即SP算法)。对两种算法进行了分析和比较,并在此基础上提出了一种改进算法。通过实验验证所改进的分类器是正确的、有效的。An augmented naive Bayes classifier of Bayes classifier family is studied in this paper. This classifier is defined by the following two conditions-one is that each attribute has the class attribute as parent;the other is that each attribute may have one other attribute as parent. Both representative algorithms are Hill Climbing Search and Superparent. Analysis and comparison are done to the above two algorithms,proposing an improved algorithm at the same time. It is sure that the modified algorithm is effective and correct during the demonstration.

关 键 词：朴素贝叶斯分类器 贪婪爬山算法 超父结点算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]