期刊论文

SVM Based Missing Data Imputation Algorithm in Nuclear Power Plant＇s Environmental Radiation Monitor Sensor Network

中文

南华大学学报(自然科学版)

1673-0062

2012

26

4

14-17

10.19431/j.cnki.1673-0062.2012.04.004

湖南省科技厅科研基金资助项目（2012FJ4332）；

本校为第一机构

传感器节点监测数据缺失会影响核电站外围环境辐射监测的有效性,需要对缺失数据进行准确估计.提出一种基于支持向量机的监测数据缺失值估计算法,对传感器节点缺失监测数据进行估计.用实际监测数据对算法进行了验证,用均方误差和相关系数评价实验结果.并与现有的基于神经网络的估计算法进行了性能比较.实验结果表明,本文所提出的算法具有较高的估计精度.

Monitor data mission in sensor node will influence the validity of environmental ra- diation monitoring around a nuclear power plant. To solve the problem, a missing data imputa- tion algorithm based on support vector machine（SVM） is proposed to impute the missing data. This Algorithm is validated with real radiation monitoring data, evaluated by using MSE and correlation coefficient and compared with ANN based imputation algorithm. The experimental results demonstrate that the ...