论文部分内容阅读
将独立成分分析用于紫外光谱定量分析,结合多模型共识的基本思想,建立了共识独立成分回归方法。从训练集随机取样建立一系列独立成分回归模型,选取其中性能较好的部分模型作为成员模型,并用这些成员模型预测未知样品。用该方法对苯甲酸、苯胺及苯酚3组分水溶液的紫外光谱进行分析,并与单模型偏最小二乘法了进行比较。结果PLSR对独立测试集中3种组分进行50次重复预测的平均RMSEP分别为2.349,7.413和1.605,RMSEP的标准偏差分别为1.781,2.918和1.266;而本方法重复50次预测的平均RMSEP分别为1.633,3.390和1.496,RMSEP的标准偏差分别为6.642×10~(-3),6.573×10~(-2)和4.484×10~(-2)。可见,共识独立成分回归所建立的模型更加稳健和可靠,预测的准确性也明显提高。
The independent component analysis is applied to the quantitative analysis of ultraviolet spectrum. Combined with the basic idea of multi-model consensus, a consensus independent component regression method is established. A series of independent component regression models were randomly sampled from the training set to select part of the models with good performance as the member models and predict the unknown samples with these member models. The UV spectra of the aqueous solutions of benzoic acid, aniline and phenol were analyzed by this method and compared with the single-model partial least squares method. Results The average RMSEP of 50 repeated predictions of PLSR for the three independent components was 2.349, 7.413 and 1.605, respectively, and the standard deviations of RMSEP were 1.781, 2.918 and 1.266, respectively. The average RMSEP of PLSR over 50 predictions was The standard deviations of RMSEP were 6.642 × 10 ~ (-3), 6.573 × 10 ~ (-2) and 4.484 × 10 ~ (-2), respectively. It can be seen that the model established by the regression of independent components of consensus is more robust and reliable, and the accuracy of forecasting is obviously improved.