陈静,蒋正凯,付敬奇.基于Netica的自学习贝叶斯网络的构建[J].电子测量与仪器学报,2016,30(11):1687-1693
基于Netica的自学习贝叶斯网络的构建
Construction of self learning Bayesian network based on Netica
  
DOI:10.13382/j.jemi.2016.11.009
中文关键词:  贝叶斯网络  网络学习  概率推理  证据敏感性分析
英文关键词:Bayesian network  learning network  probability reference  sensitivity analysis of finding
基金项目:国家“863”高技术研究发展计划(2001AA040103 7) 资助项目
作者单位
陈静 安徽理工大学 电气学院淮南232001 
蒋正凯 安徽理工大学 电气学院淮南232001 
付敬奇 上海大学机自学院上海200072 
AuthorInstitution
Chen Jing School of Electrical and Information Engineering, Anhui University of Science and Technology, Huainan 232001, China 
Jiang Zhengkai School of Electrical and Information Engineering, Anhui University of Science and Technology, Huainan 232001, China 
Fu Jingqi School of Mechatronics Engineering and Automation, Shanghai University, Shanghai 200072, China 
摘要点击次数: 141
全文下载次数: 155
中文摘要:
      针对贝叶斯网络构建时参数与结构难以自适应调整,提出基于Netica的自学习贝叶斯网络的构建方法。首先根据Netica要求处理样本数据集,然后运用Netica基础函数开发结构学习模块和参数学习模块,进而能够构建出自动学习样本数据集的贝叶斯网络。同时,开发了概率推理模块和证据敏感性分析模块以评估所建网络的有效性。以国家电网的短路故障样本数据为例建立其自学习贝叶斯网络,实验构建的自学习贝叶斯网络能够实现不确定性推理,表明所提方法是贝叶斯网络功能实现的一个新途径。
英文摘要:
      Aiming at the limitation of adaptive adjustment of parameter and structure in process of constructing Bayesian network, a construction way of self learning Bayesian network based on Netica software is presented. Firstly, the sample data set is processed according to the request of Netica. Then, structure learning module and parameter learning module is developed by using Netica basis function, in turn. Bayesian network is established by learning sample data set automatically. At the same time, probability references module and evidence sensitivity analysis module are developed to assess the effectiveness of the built network. The self learning Bayesian network is established by using short circuit fault sample data set from state grid, and the self learning Bayesian network can implement the uncertainty reference. The result shows that the self learning Bayesian network is feasible and effective, and it provides a new way for the realization of the Bayesian network.
查看全文  查看/发表评论  下载PDF阅读器