近红外多组分分析中异常样本识别方法

Outlier Samples Detection Method for NIR Multicomponent Analysis

近红外光谱分析中,异常样本的存在严重影响定标模型的预测性能和适配性。基于X/Y联合的ODXY异常样本识别和剔除方法,提出并证明了一种专用于多组分分析的MODXY异常样本识别方法。实验采用80组玉米近红外光谱数据,利用不同异常样本识别方法剔除异常样本后建立玉米含水率、含油率、蛋白质含量和淀粉含量4种组分的偏最小二乘预测模型,采用预测均方差和决定系数作为评价指标比较所建模型的性能,检验MODXY方法在多组分分析中的异常样本识别能力。实验结果表明:在近红外多组分分析中,MODXY方法在大多数情况下具有更好的异常样本识别能力;MODXY方法和ODXY方法均有一定的适用范围,它们更适合于相应组分化学值的相对标准偏差较大的情况。

Near infrared spectroscopy is currently a highly versatile tool used in diverse fields. However,outlier samples strongly affect the performance of the prediction model in near infrared spectroscopy analysis. Therefore,detecting and eliminating the outlier samples is a major and important procedure in near infrared spectroscopy analysis. Using the outlier samples detection based on joint X-Y distances( ODXY) method,a special outlier samples detection method for multicomponent analysis was proposed and proved,termed as MODXY method. Experimental data was derived from the near infrared spectra of80 corns. Based on these,the PLS models of moisture content,oil content,protein content and starch content were constructed by eliminating outlier samples using different outlier detection methods. The obtained models were compared in terms of performance by the predictive root mean square error( RMSEP) and the coefficient of determination( R2). The results showed that in most cases the MODXY method had better outlier sample recognition capability in NIR multicomponent analysis compared with other methods. Both ODXY method and MODXY method had their own suitable range,and they were more effective when the relative standard deviation of components was large enough.