完全贝叶斯分类器在经济数据分类中的应用  

Application of complete bias classifier in economic data classification

作　　者：王辉[1] 李玉亮 王莉[1] WANG Hui;LI Yu-liang;WANG Li(School of Information Engineering,Minzu University of China,Beijing 100081,China)

机构地区：[1]中央民族大学信息工程学院

出　　处：《东北师大学报（自然科学版）》2019年第4期65-69,共5页Journal of Northeast Normal University(Natural Science Edition)

基　　金：国家自然科学基金资助项目（61672553）;教育部社科基金资助项目（18YJAZH087)

摘　　要：针对贝叶斯分类器分类强关联属性导致分类准确率下降的问题,提一种完全贝叶斯分类器合理利用属性间的依赖关系优化贝叶斯分类器,对参数进行动态调整组合,同时合理剔除无关属性.采用国内外知名数据库提供的数据,通过与其他分类器的对比实验,证明了完全贝叶斯分类器在宏观与微观经济数据分类中都获得了较好的分类效果.Bayesian classifier has the advantages of high classification stability and simple algorithm implementation,so it is widely used at present.However,Bayesian classifier can t solve the problem that the classification accuracy is reduced due to the strong correlation between multiple attributes.In this paper,a complete Bayesian classifier is explored to optimize the Bayesian classifier by making use of the dependency among attributes reasonably.In the process of implementation,parameters are dynamically adjusted and combined,and irrelevant attributes are reasonably removed.In the experiment,we use the data provided by the well-known databases at home and abroad,and through the contrast experiment with other classifiers,it is proved that the complete Bayesian classifier has better classification results in the macro and micro economic data classification.

关 键 词：数据挖掘 完全贝叶斯分类器 半朴素贝叶斯分类器 分类 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]