一种模态傅里叶-支持向量机优化的取用水监测异常数据重构方法
|
摘要
提高数据的完备与真实性是水资源监控能力建设的关键。针对国家水资源监控能力建设项目实施以来其监测数据呈现出的异常特征,按照"先粗筛后精选"逻辑,并考虑取用水季节性周期波动的特点,提出采用拉依达准则-模态分解-傅里叶残差修正的水监测数据异常值识别方法,并根据粒子群优化最小二乘支持向量机模型实现对异常数据的重构恢复。通过对企业取用水数据的实例分析,结果表明分段式拉依达准则在其监测异常数据的粗筛中具有较好的适用性,利用傅里叶修正集合模态分解的监测数据序列可取得更佳的拟合效果,从而达到异常数据精选的目的;而粒子群优化最小二乘支持向量机模型对异常数据重构恢复的可信度高于普通最小二乘支持向量机及传统曲线拟合数据重构方法,即该类取用水监测异常数据重构方法可有助于进一步推进其监测数据对实际水资源状态的客观反映。
Data completeness and authenticity to be improved is the key of water resources monitoring capacity building project. In this paper,according to the abnormal characteristics of water resources monitoring data in the project and the seasonal cycle fluctuation rule of industrial water,methods of abnormal data detection were proposed,which including Pauta criterion,ensemble empirical mode decomposition( EEMD) and Fourier function. After that,the abnormal data could be reconstructed by particle swarm optimization least squares support vector machine( PSO-LSSVM) model. All of above methods were tested empirically in the water resources monitoring data of industrial company. Results showed that Pauta criterion had a good applicability in the preliminary judgment of abnormal data when it was used by sectionalized method. Moreover,although monitoring data could be fitted,the fitting residuals needed to be fixed,and it was realized by Fourier function so that the fitting effect of monitoring data was better. In terms of data reconstruction and recovery,PSO-LSSVM model had higher credibility than LSSVM and traditional curve fitting method. Hence,all of these methods of abnormal data reconstruction could be well applied in the national water resource information monitoring system in China.
Data completeness and authenticity to be improved is the key of water resources monitoring capacity building project. In this paper,according to the abnormal characteristics of water resources monitoring data in the project and the seasonal cycle fluctuation rule of industrial water,methods of abnormal data detection were proposed,which including Pauta criterion,ensemble empirical mode decomposition( EEMD) and Fourier function. After that,the abnormal data could be reconstructed by particle swarm optimization least squares support vector machine( PSO-LSSVM) model. All of above methods were tested empirically in the water resources monitoring data of industrial company. Results showed that Pauta criterion had a good applicability in the preliminary judgment of abnormal data when it was used by sectionalized method. Moreover,although monitoring data could be fitted,the fitting residuals needed to be fixed,and it was realized by Fourier function so that the fitting effect of monitoring data was better. In terms of data reconstruction and recovery,PSO-LSSVM model had higher credibility than LSSVM and traditional curve fitting method. Hence,all of these methods of abnormal data reconstruction could be well applied in the national water resource information monitoring system in China.
引文
[1]中华人民共和国水利部,财政部.国家水资源监控能力建设项目实施方案(2012-2014)[S].北京:中华人民共和国水利部,2012.
[2] Kurtz W,Lapin A,Schilling O S,et al. Integrating hydrological modelling,data assimilation and cloud computing for real-time management of water resources[J].Environmental Modelling&Software, 2017(93):418-435.
[3] Mun珘iz C D,Nieto P J G,Fernández J R A,et al. Detection of outliers in water quality monitoring samples using functional data analysis in San Esteban estuary(Northern Spain)[J]. Science of the Total Environment,2012(439):54-61.
[4] Di Blasi J I P,Torres J M,Nieto P J G,et al. Analysis and detection of outliers in water quality parameters from different automated monitoring stations in the Min珘o river basin(NW Spain)[J]. Ecological Engineering,2013(60):60-66.
[5] Shirdel A H,Bling J M,Toivonen H T. System identification in the presence of trends and outliers using sparse optimization[J]. Journal of Process Control,2016(44):120-133.
[6] Aljoumani B,Sànchez-Espigares J A,Can珘ameras N,et al. Time series outlier and intervention analysis:irrigation management influences on soil water content in silty loam soil[J]. Agricultural Water Management,2012(111):105-114.
[7]龙秋波,贾绍凤,汪党献.中国用水数据统计差异分析[J].资源科学,2016,38(2):248-254.
[8]毛莺池,齐海,接青,等. M-TAEDA:多变量水质参数时序数据异常事件检测算法[J].计算机应用,2017,37(1):138-144.
[9]方海泉,薛惠锋,蒋云钟,等.基于EEMD的水资源监测数据异常值检测与校正[J].农业机械学报,2017,48(9):257-263.
[10]魏晶茹,马瑜,白冰,等.基于PSO-SVM算法的环境监测数据异常检测和缺失补全[J].环境监测管理与技术,2016,28(4):53-56.
[11] Tyanova S,Temu T,Sinitcyn P,et al. The perseus computational platform for comprehensive analysis of(prote)omics data[J]. Nature Methods,2016,13(9):731-740.
[12] Huang N E,Shen Z,Long S R,et al. The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis[C]//Proceedings of the Royal Society of London A:Mathematical,Physical and Engineering Sciences. The Royal Society,1998,454:903-995.
[13] Wu Z,Huang N E. Ensemble empirical mode decomposition:a noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[14]黄元生,方伟.基于灰色傅里叶变换残差修正的电力负荷预测模型[J].电力自动化设备,2013,23(9):105-107.
[15] Salehi E,Abdi J,Aliei M H. Assessment of Cu(II)adsorption from water on modified membrane adsorbents using LS-SVM intelligent approach[J]. Journal of Saudi Chemical Society,2016,20(2):213-219.
[16] Jahn J. Karush-kuhn-tucker conditions in set optimization[J]. Journal of Optimization Theory and Applications,2017,172(3):707-725.
[17] García-Valenzuela A,Márquez-Islas R,Barrera R G.Reducing light-scattering losses in nanocolloids by increasing average inter-particle distance[J]. Applied Physics A,2017,123(1):80-84.
[18] Rai P K,Dhanya C T,Chahar B R. A PSO approach for optimum design of dynamic inversion controller in water distribution systems[J]. Journal of Water Supply:Research and Technology-Aqua, 2016, 65(7):570-581.
[19] Etienne L,Devogele T,Buchin M,et al. Trajectory box plot:a new pattern to summarize movements[J]. International Journal of Geographical Information Science,2016,30(5):835-853.
[2] Kurtz W,Lapin A,Schilling O S,et al. Integrating hydrological modelling,data assimilation and cloud computing for real-time management of water resources[J].Environmental Modelling&Software, 2017(93):418-435.
[3] Mun珘iz C D,Nieto P J G,Fernández J R A,et al. Detection of outliers in water quality monitoring samples using functional data analysis in San Esteban estuary(Northern Spain)[J]. Science of the Total Environment,2012(439):54-61.
[4] Di Blasi J I P,Torres J M,Nieto P J G,et al. Analysis and detection of outliers in water quality parameters from different automated monitoring stations in the Min珘o river basin(NW Spain)[J]. Ecological Engineering,2013(60):60-66.
[5] Shirdel A H,Bling J M,Toivonen H T. System identification in the presence of trends and outliers using sparse optimization[J]. Journal of Process Control,2016(44):120-133.
[6] Aljoumani B,Sànchez-Espigares J A,Can珘ameras N,et al. Time series outlier and intervention analysis:irrigation management influences on soil water content in silty loam soil[J]. Agricultural Water Management,2012(111):105-114.
[7]龙秋波,贾绍凤,汪党献.中国用水数据统计差异分析[J].资源科学,2016,38(2):248-254.
[8]毛莺池,齐海,接青,等. M-TAEDA:多变量水质参数时序数据异常事件检测算法[J].计算机应用,2017,37(1):138-144.
[9]方海泉,薛惠锋,蒋云钟,等.基于EEMD的水资源监测数据异常值检测与校正[J].农业机械学报,2017,48(9):257-263.
[10]魏晶茹,马瑜,白冰,等.基于PSO-SVM算法的环境监测数据异常检测和缺失补全[J].环境监测管理与技术,2016,28(4):53-56.
[11] Tyanova S,Temu T,Sinitcyn P,et al. The perseus computational platform for comprehensive analysis of(prote)omics data[J]. Nature Methods,2016,13(9):731-740.
[12] Huang N E,Shen Z,Long S R,et al. The empirical mode decomposition and the hilbert spectrum for nonlinear and non-stationary time series analysis[C]//Proceedings of the Royal Society of London A:Mathematical,Physical and Engineering Sciences. The Royal Society,1998,454:903-995.
[13] Wu Z,Huang N E. Ensemble empirical mode decomposition:a noise-assisted data analysis method[J].Advances in Adaptive Data Analysis,2009,1(1):1-41.
[14]黄元生,方伟.基于灰色傅里叶变换残差修正的电力负荷预测模型[J].电力自动化设备,2013,23(9):105-107.
[15] Salehi E,Abdi J,Aliei M H. Assessment of Cu(II)adsorption from water on modified membrane adsorbents using LS-SVM intelligent approach[J]. Journal of Saudi Chemical Society,2016,20(2):213-219.
[16] Jahn J. Karush-kuhn-tucker conditions in set optimization[J]. Journal of Optimization Theory and Applications,2017,172(3):707-725.
[17] García-Valenzuela A,Márquez-Islas R,Barrera R G.Reducing light-scattering losses in nanocolloids by increasing average inter-particle distance[J]. Applied Physics A,2017,123(1):80-84.
[18] Rai P K,Dhanya C T,Chahar B R. A PSO approach for optimum design of dynamic inversion controller in water distribution systems[J]. Journal of Water Supply:Research and Technology-Aqua, 2016, 65(7):570-581.
[19] Etienne L,Devogele T,Buchin M,et al. Trajectory box plot:a new pattern to summarize movements[J]. International Journal of Geographical Information Science,2016,30(5):835-853.