论文部分内容阅读
利用近红外光谱分析结合偏最小二乘法测定甲苯、氯苯和正庚烷构成的三元混合体系中氯苯的浓度。通过考察混合物吸光度与氯苯浓度的相关性,以预测残差平方和为评价指标,对光谱进行了变量筛选,并将选择的33个波长变量建立的定量校正模型与全谱范围的校正模型进行比较。结果表明本方法可使最终所得的模型中变量数大大减少,通过波长变量优选最小化冗余信息、提高预测精度及建模效率。
The concentration of chlorobenzene in the ternary mixed system consisting of toluene, chlorobenzene and n-heptane was determined by near infrared spectroscopy combined with partial least squares. By investigating the correlation between the absorbance of the mixture and the concentration of chlorobenzene, the square of the residual was used as the evaluation index, the spectrum was screened, and the quantitative calibration model established by the selected 33 wavelength variables and the full-scale calibration model Compare The results show that this method can greatly reduce the number of variables in the final model, and minimize the redundant information by using wavelength variables to improve the prediction accuracy and modeling efficiency.