A dual model strategy to transfer multivariate calibration models for near-infrared spectral analysis. Issue 5 (27th May 2016)

• Record Type: Journal Article
• Authors: Wang, Jiajun; Li, Zhengfeng; Wang, Yi; Liu, Yan; Cai, Wensheng; Shao, Xueguang
• Is Part Of: Spectroscopy letters. Volume 49:Issue 5(2016)
• Page Start: 348
• Page End: 354
• Publication Date: 2016-05-27
• Subjects: Calibration model transfer -- correction model -- dual-model strategy -- near-infrared spectroscopy; Spectrum analysis -- Periodicals; 543.5
• DOI: 10.1080/00387010.2016.1167087
• Languages: English
• ISSNs: 0038-7010

Abstract

ABSTRACT: A dual-model method is proposed for correcting the calibration model. In the method, a primary calibration model is built using the spectra of a primary instrument and a correction model is established to describe the ratios between the predicted results from the spectra of different instruments. The prediction for the spectra of secondary instrument can be achieved by correcting the prediction of the primary model. A mathematical proof is described for the existence of the correction model, and the model is investigated using a near-infrared spectroscopic dataset of plant leaf samples measured on two instruments. The results show that a precise correction model is obtained and the model can be used to correct the predictions of the primary model. The correlation coefficients between the predicted and the reference ratios are above 0.9, and the prediction error after the correction is at the same level of the primary model.