摘要
测量了含微量甲醇(体积分数为0.04%～0.24%)的系列乙醇水溶液的近红外光谱,利用近红外光谱分析建立了预测甲醇含量的定量分析模型。比较了用外部检验法(Test Set-Validation)和交叉检验法(Cross-Validaton)建立的数学模型,研究了使用外部检验法时,校正集和检验集样品数的改变对模型预测结果的影响。结果发现,当校正集样品数为15检验集样品数为6(总样品数为21)时,使用外部检验法建立的数学模型预测结果较好,其校正集的均方根误差和检验集的预测均方根误差(分别为RMSEE和RMSEP)均较小(分别为0.0115和0.0105),而且很接近。结果表明,近红外光谱方法简单,准确而且实用。
Near - infrared spectra of ethanol solutions were obtained and the mathematic models for quantitative analysis of trace methanol in solution were set up. The mathematic models established by test set-validation and by cross-validation were compared, and the effect of the change of the numbers of samples used in calibration model and validation model of test set-validation on the mathematic models was evaluated. It was found that the predicted results of the mathematic model are good when the number of samples is 15 in calibration model and that in validation model is 6 （ the total is 21 ）. In that case, the value of the root mean square error of the estimation of calibration and the root mean square error of prediction of validation （ RMSEE and RMSEP, respectively） are very small（0. 011 5 and 0. 010 5 respectively） and almost the same. Our results indicate that near infrared spectroscopy can be used to analyze the content of methanol in ethanol solution very well.
出处
《应用化学》
CAS
CSCD
北大核心
2007年第10期1101-1104,共4页
Chinese Journal of Applied Chemistry
基金
国家自然科学基金资助项目(20273068
90306001)
关键词
近红外光谱
数学模型
定量分析
甲醇
near-infrared spectroscopy, mathematic model, quantitative analysis, methanol
