Journal
SPECTROCHIMICA ACTA PART A-MOLECULAR AND BIOMOLECULAR SPECTROSCOPY
Volume 205, Issue -, Pages 227-234Publisher
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.saa.2018.07.029
Keywords
Congou black tea fermentation; Theaflavin-to-thearubigin ratio; NIR spectroscopy; Extreme learning machine; Variable selection
Categories
Funding
- National Key R&D Program of China [2018YFD0700500, 2017YFD0400802]
- Natural Science Foundation of Zhejiang Province [Y16C160009]
- Central Public-interest Scientific Institution Basal Research Fund [1610212018005]
- Key R&D Program of Zhejiang Province [2515C02001]
- Innovation Project of Chinese Academy of Agricultural Sciences [CAAS-ASTIP-2014-TRICAAS]
- China Agriculture Research System [CARS-23]
Ask authors/readers for more resources
The theaflavin-to-thearubigin ratio (TF/TR) is an important parameter for evaluating the degree of fermentation and quality characteristics of Congou black tea. Near infrared (NIR) spectroscopy, one of the most promising techniques for evaluating large-scale tea processing quality, in association with chemometrics, can be used as a selection tool when a fast determination of the requested parameters is required. The aim of this work is to develop a unique model for the determination of TF/TR. First, 11 key wavelength variables were screened by synergy interval partial least-squares regression (SI-PLS) and competitive adaptive reweighted sampling (CARS). Based on these characteristic variables, a new extreme learning machine (ELM) combined with an adaptive boosting (ADABOOST) algorithm (ELM-ADABOOST) was applied to construct the nonlinear prediction model for TF/TR, and an independent external set was used for the validation. A determinate coefficient (R-P(2)) of 0.893, root mean square error of prediction (RMSEP) of 0.0044, RSD below 10%, and RPD above 3 were acquired in the prediction model. These results demonstrate that NIR can be used to rapidly determine the TF/TR value during fermentation, and it effectively simplify the model and improve the prediction accuracy when combined with the SI-CARS variable. (C) 2018 Elsevier B.V. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available