华中科技大学学报（自然科学版）Issue(z1)：81-83,87,4.DOI:10.13245/j.hust.15S1020
基于数据扩展的动态贝叶斯网络预测方法
Prediction method for dynamic Bayesian network based on data extension
摘要
Abstract
Data extension method was applied to the dynamic Bayesian network for parameter learn-ing,using random sampling algorithm to extend the data of small samples and the Bayesian posterior probability formula to modify extended data.In the meantime,the posterior probability of the ob-served data was calculated,and then on the basis of the extended data,learning and reasoning of the dynamic Bayesian network were completed.Simulation results show that this method can reduce error accumulation caused by the combined effects of nodes in the prediction model so as to improve the ac-curacy of prediction of the model.关键词
动态贝叶斯网络/误差积累/数据扩展/条件概率/故障预测Key words
dynamic Bayesian network/error accumulation/data extension/conditional probability/fault prediction分类
信息技术与安全科学引用本文复制引用
刘春阳,张泽浩,柳长安,吴华..基于数据扩展的动态贝叶斯网络预测方法[J].华中科技大学学报（自然科学版）,2015,(z1):81-83,87,4.基金项目
国家自然科学基金资助项目（61105083）；新世纪优秀人才资助计划资助项目（NCET-11-0634）；北京市教委资助项目（GJ2013005）． （）
