基于紫外吸收光谱的水质硝酸盐含量LLE-SVR建模  

LLE-SVR modeling for nitrate measurement in water based on UV absorption spectra

作　　者：潘国锋[1] 

机构地区：[1]江南大学轻工过程先进控制教育部重点实验室,无锡214122

出　　处：《仪器仪表学报》2011年第12期2869-2873,共5页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60674092);无锡市污染防治基金(2008-1)资助项目

摘　　要：基于紫外吸收光谱的总氮测定方法在分析精度上严重依赖于所建数学模型的预测精度。针对紫外光谱数据波段多,容易引入未知物质干扰光谱信息的问题,提出了局部线性嵌入(local linear embedded,LLE)和偏最小二乘支持向量回归机(least square support vector regression machine,LS-SVR)相结合的紫外吸收光谱数据建模方法。首先利用LLE算法将高维的紫外吸收光谱映射到低维的流形空间,实现高维非线性光谱数据结构的特征提取,并利用LS-SVR建立硝酸盐含量的非线性回归模型。仿真结果表明:利用LLE-SVR方法获得的硝酸盐含量预测模型,其训练样本的相对误差为0.001 9,测试样本相对误差为0.035 8,小于单纯的LS-SVR模型的0.023 3和0.060 2。The analysis accuracy of nitrogen measurement based on UV absorption spectrum method depends on the prediction accuracy of the mathematic model. A local linear embedded （LLE） method coupled with least square support vector regression （LS-SVR） modeling method is proposed for the mass data of UV spectra to reduce the interference from unknown substances. First, LLE is used to reduce the dimensions of UV absorption spectra, and the low-dimensional features are extracted. Then, a nonlinear regression model is developed using LS-SVR method to predict the nitrate concentration. Simulation results show that the relative error of the training set is 0. 001 9 and that of the prediction set is 0. 038 5, which are lower than those of the single SVR model ： 0. 022 3 for the training set and 0. 062 for the prediction set.

关 键 词：硝酸盐 支持向量回归机 紫外吸收光谱 局部线性嵌入 模型精度 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置] O657.32[自动化与计算机技术—控制科学与工程]