应用典型相关分析法校正不同仪器和不同时间测量的近红外光谱(英文)  

Application of canonical correlation analysis(CCA) to correct NIR spectra measured on different instruments and in different measurement dates

作　　者：郑开逸[1] 范伟[1,2] 吴婷[1] 胡慧廉[1] 杜一平[1] 梁逸曾[2] 

机构地区：[1]上海市功能性材料化学重点实验室,华东理工大学分析测试中心,上海200237 [2]中药现代化研究中心,中南大学化学化工学院,长沙湖南410083

出　　处：《计算机与应用化学》2013年第3期246-250,共5页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(20975039,21205041)

摘　　要：以甲苯、氯苯和正庚烷混和三组分体系为分析样品,应用模型转移法来校正不同仪器和不同测量时间导致的近红外光谱之间的差异。在校正测量时间导致的差异时,以16日由Bruker近红外光谱仪扫描的光谱作为主光谱,而17日由同台仪器扫描的光谱作为从光谱。在校正仪器导致的光谱差异时,以16日由Bruker近红外光谱仪扫描的光谱作为主光谱,而在同一天由Perkin-Elmer近红外光谱仪扫描的光谱作为从光谱。采集光谱后,用分段直接校正(PDS)和典型相关分析(CCA)校正从光谱。结果显示:虽然仪器和测量时间差别均导致从光谱偏离主光谱,但是CCA的效果更佳,使得校正后的从光谱能用主光谱的模型准确预测样品含量。Calibration transfer methods were used to correct near infrared spectra （NIR） of tri-component mixtures including methylbenzene, chlorobenzene and n-heptane for their variations resulting from differences of measurement date and instrument. For correcting spectra between different measurement dates and instruments the spectra of mixtures measured in spectrometer of Bruker on date 16 were assigned as primary spectra while the spectra of same mixtures detected in spectrometer of Bruker on date 17 and in spectrometer of Perkin-Elmer on date 16 were set as secondary spectra for changes of measurement date and instrument respectively. After spectra obtained, two calibration transfer methods,including canonical correlation analysis （CCA） and piecewise direct standardization （PDS） were used to correct the secondary spectra. The results showed that although variations of both measurement date and instrument can cause secondary spectra to deviate from primary spectra, the CCA can correct the secondary spectra to use the model of primary spectra to predict those three components with high accuracy.

关 键 词：三组分体系 模型转移 CCA PDS 测定时间 近红外光谱仪 

分 类 号：TQ021.4[化学工程]