基于广义朴素贝叶斯分类器的空值处理方法  

Research on the method of processing empty value based on generalized naive Bayes classifiers

作　　者：王辉[1] 王双成[2] 周颜军[1] 张剑飞[1] 

机构地区：[1]东北师范大学计算机科学系,吉林长春130024 [2]吉林大学计算机科学与技术系,吉林长春130025

出　　处：《东北师大学报（自然科学版）》2004年第1期34-38,共5页Journal of Northeast Normal University(Natural Science Edition)

基　　金：吉林省自然科学基金资助项目(20030517-1)

摘　　要：　给出了基于广义朴素贝叶斯分类器的空值处理方法,该方法利用广义朴素贝叶斯分类器的较强预测能力及灵活的效率选择方式,有效地解决了数据中的空值处理问题.利用模拟数据进行了对比实验,结果显示,其预测准确性明显提高.In this paper,the method of processing empty value based on generalized naive Bayes classifiers is developed,the method can make use of the high accurate rate in prediction and flexibility in choosing efficiency of generalized naive Bayes classifiers to dispose the problem with empty value effectively.Meantime,a contrast experiment is made by simulated data and the results of the experiment is given,the accurate rate of prediction is improved.

关 键 词：广义朴素贝叶斯分类器 空值处理 分类能力 

分 类 号：TP311.134.1[自动化与计算机技术—计算机软件与理论]