多维数据分析法同时测定伏安峰严重重叠的有机污染物

Simultaneous Determination of Three Organic Pollutants with Severe Over-Lapping Volt-Ampere Peaks by Multi-Dimensional Data Analysis

设计了一种正交信号校正(OSC)-小波包变换(WPT)-偏最小二乘法(PLS)的多维数据分析方法,该方法结合微分脉冲伏安技术可对不经化学分离同时伏安峰严重重叠的2,5-二硝基酚、间硝基酚和对硝基酚的混合物进行同时测定。该方法利用OSC有效除去与浓度无关的结构噪音,又利用WPT改善了除噪和特征信息提取能力,从而提高了PLS的预测能力。根据OSC,WPT,PLS的算法原理编制了名为POSCWPTPLS,PWPTPLS,PPLS的程序来执行相关的计算。OSC-WPT-PLS,WPT-PLS,PLS法的测定结果表明,3种化合物的总相对预测标准偏差分别为3.76%,7.39%,7.98%,OSC-WPT-PLS法的结果优于WPT-PLS法和PLS法。

A multi-dimensional data analysis method named OSC-WPT-PLS based on partial least square（PLS） regression with orthogonal signal correction（OSC） and wavelet packet transform（WPT） as preprocessing tools, which was combined with different pulse voltammetry, was proposed for simultaneous determination of 2,5-dinitrophenol, 3-nitrophenol and 4-nitrophenol in a mixture with severe overlapping volt-ampere peaks without prior chemical separation. OSC was used to remove the structure noise that was irrelative to concentrations of the compounds. WPT was used to perform data compression, to extract relevant information and to eliminate noise. PLS was applied for multivariate calibration and noise reduction by eliminating the less important variables. Three programs （POSCWPTPLS,PWPTPLS and PPLS ）were designed to carry out the relative calculation. Results showed the relative standard errors of prediction for all components with OSC-WPT-PLS, WPT-PLS and PLS were 3.76%, 7.39% and 7.98%, respectively. Compared with the other two methods, OSC-WPT-PLS method combines all the advantages from OSC-WPT and PLS for enhancing the ability in the extraction of character information and the quality of regression, and its result is satisfactory.