PCA-based algorithm for calibration of spectrophotometric analysers of food

2013 Joint IMEKO (International Measurement Confederation) TC1-TC7-TC13 Symposium: Measurement Across Physical and Behavioural Sciences

Morawski, Roman Z.^1 ; Mikina, Andrzej^1

Warsaw University of Technology, Faculty of Electronics and Information Technology, Institute of Radioelectronics, Warsaw, Poland^1

关键词: Food industries;    Metrological performance;    Numerical performance;    Principal components analysis;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/459/1/012029/pdf
DOI  :  10.1088/1742-6596/459/1/012029

来源: IOP

【 摘 要 】

Spectrophotometric analysers of food, being instruments for determination of the composition of food products and ingredients, are today of growing importance for food industry, as well as for food distributors and consumers. Their metrological performance significantly depends of the numerical performance of available means for spectrophotometric data processing; in particular-the means for calibration of analysers. In this paper, a new algorithm for this purpose is proposed, viz. the algorithm using principal components analysis (PCA). It is almost as efficient as PLS-based algorithms of calibration, but much simpler.