基于贝叶斯置信区间的传感器网络故障诊断方法研究  

Study of Fault Diagnosis for Sensor Network Based on Confidence Intervals of Bayesian

作　　者：何永强[1] 秦勤[1] 

机构地区：[1]河南工程学院计算机学院,郑州451191

出　　处：《重庆科技学院学报（自然科学版）》2016年第5期90-93,共4页Journal of Chongqing University of Science and Technology：Natural Sciences Edition

基　　金：河南省科技厅科技计划课题"基于云模型和模糊数学的无线传感器网络鲁棒性研究"(152102210027);河南省高等学校重点科研项目"基于贝叶斯的传感器网络故障检测研究"(15A520054)

摘　　要：通过构建贝叶斯网络故障诊断模型,提出了结合故障诊断准确性和置信区间对传感器网络进行故障诊断的方法。采用二项分布参数估计方法,结合β分布和F分布对置信区间进行算法改进,降低了对测试数据规模的要求,减小了准确度的误差。通过仿真分析,验证了改进算法的准确性和可靠性。This paper constructs fault diagnosis model of Bayesian network and presents the method from sensor net- work fault diagnosis accuracy and confidence intervals to carry on fault diagnosis of sensor networks. Algorithms for confidence intervals are improved based on Binomial distribution combined with - distribution and - distribution to reduce requirement for scale of the test data and decrease error of accuracy. Through simulation analysis, the accu- racy and reliability of the improved algorithm is verified.

关 键 词：贝叶斯 置信区间 故障诊断 传感器网络 准确度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]