用平滑方法改进多关系朴素贝叶斯分类.pdf

Computer Engineering and Applications 计算机工程与应用 2017 ，53（5） 69 ⦾大数据与云计算⦾ 用平滑方法改进多关系朴素贝叶斯分类 1 2 1 1 徐光美 ，刘宏哲 ，张敬尊 ，王金华 1 2 1 1 XU Guangmei , LIU Hongzhe , ZHANG Jingzun , WANG Jinhua 1.北京联合大学 信息学院，北京100101 2.北京联合大学 信息服务工程重点实验室，北京 100101 1. College of Information Technology, Beijing Union University, Beijing 100101, China 2. Beijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing 100101, China XU Guangmei, LIU Hongzhe, ZHANG Jingzun, et al. Improving multi-relational Naive Bayesian classifier using smoothing methods. Computer Engineering and Applications, 2017, 53 （5 ）：69-72. Abstract: To eliminate the naive Bayesian classification of zero probability and overfitting problem, this paper discusses the various probability smoothing method, gives MRNBC-M （Multi-Relational Naive Bayesian Classifier based on M- estimation ）and MRNBC- L （Multi- Relational Naive Bayesian Classifier based on Laplace- estimation ）. In the case of multi-relationship, the impact of M and Laplace estimation methods on the classification is analyzed. In order to further optimize the classification, the method is based on the extended mutual information criterion. Experiments on the multi- relational datasets show that MRNBC-M can effectively improve the classification performance. Key words: Multi-Relational Data Mining （MRDM ）; Naive Bayes; smoothing methods; mutual information 摘 要：为消除朴素贝叶斯分类时的零概率以及过度拟合问题，分析了各种概率平滑方法，给出了基于M 估计的多 关系朴素贝叶斯分类方法（MRNBC-M ）和基于Laplace 估计的多关系朴素贝叶斯分类方法（MRNBC-L ），分析探讨了 M 平滑和Laplace 平滑方法对多关系分类的影响情况，为进一步优化分类，方法基于扩展互信息标准对数据进行属 性过滤。多关系标准数据集上的实验显示，MRNBC-M 可以有效改进分类性能。 关键词：多关系数据挖掘；朴素贝叶斯；参数平滑；互信息 文献标志码：A 中图分类号：TP 181 doi ：10.3778/j.issn. 1002-8331.1507-0161 1 引言 多关系朴素贝叶斯分类模型（Multi-Relational Naive 朴素贝叶斯分类方法（Naive Bayesian