一种引入隐藏节点的贝叶斯网络方法A method on Bayesian Network with hidden nodes
虞欣;郑肇葆;
摘要(Abstract)：
针对简单贝叶斯分类器(NBC)中"天真"而又苛刻的条件独立假设,本文结合Fisher准则提出了一种在贝叶斯网络中引入隐藏节点的方法,用来松弛或放宽这个过于严格而且不现实的假设条件。隐藏节点的引入可以更好地描述问题,进而更好地解决贝叶斯网络在分类中的应用。实验表明,本文提出的方法可以比NBC获得更高的分类精度和更好的稳定性。在训练样本不多的情况下,平均分类精度比PCA-NBC高3%之多。
关键词(KeyWords)： 贝叶斯网络;隐藏节点;纹理;影像;分类
基金项目(Foundation): 国家自然科学基金资助项目(40571102)
作者(Authors): 虞欣;郑肇葆;
DOI: 10.16251/j.cnki.1009-2307.2011.05.019
参考文献(References)：
- [1]Yu Xin,Zheng Zhaobao,LI Linyi,Ye ZhiWei.TextureClassification of aerial image based on PCA-NBC[J].Proceedings of SPIE-The International Society for Opti-cal Engineering,v 6044,MIPPR 2005:Image AnalysisTechniques,2005.
- [2]Friedman N,Geiger D,Goldszmidt M.Bayesian Network classifiers[J].Machine Learning,1997,29:131-163.
- [3]D Heckerman,D Geiger,D M Chickering.LearningBayesian networks:the combination of knowledge and sta-tistical data[J].Machine Learning 1995,20:197-244.
- [4]Langley P,Iba W,Thompson K.An analysis of Bayesianclassifiers[J].In proceeding of AAAI-92,pp.223-228,1992.
- [5]Cheng,J,Greiner,R.Comparing Bayesian Network Clas-sifiers[Z].In UAI-99.
- [6]Yu Xin,Zheng Zhaobao,YE Zhiwe,TIAN Liqia.TextureClassification Based on Tree Augmented Naive Bayes Classifi-er[J].Geomatics And Information Science of Wuhan Uni-versity,2006.
- [7]Heckerman D.Bayesian networks for data mining[J].Data Mining and Knowledge Discovery,1997,1(1):79-119.
- [8]Friedman N.Learning belief networks in the presence ofmissing values and hidden variables[C]//Proceeding ofthe 14th International Conference on Machine Learning,San Francisco,1997:125-133.
- [9]王双成.混合贝叶斯网络隐藏变量学习研究[J].计算机学报,2005,29(9):1564-1569.
- [10]张尧庭.多元统计分析选讲[M].北京:中国统计出版社,2002.
- [11]Yu Xin,Zheng Zhaobao,Tang Ling,Ye Zhiwei.AerialImage Texture Classification Based on Naive Bayesian Net-work[J].Geomatics And Information Science of WuhanUniversity,2006,31(2).
- [12]Jiebo Luo,Andreas E.Savakis,Amit Singhal.A Bayesiannetwork-based framework for semantic image understanding[J].Pattern Recognition 2005,38:919-934.