Geographical origin of green tea identification using LASSO and ANOVA

PAN, Tianhong; YAN, Ru; CHEN, Qi

doi:10.1590/fst.41922

Acessibilidade / Reportar erro

Brasil

Food Science and Technology

Español English

Brasil

Español English

sumário « anterior atual seguinte »

Sumário

Original Article • Food Sci. Technol 42 • 2022 • https://doi.org/10.1590/fst.41922 copy

Geographical origin of green tea identification using LASSO and ANOVA

Authorship SCIMAGO INSTITUTIONS RANKINGS

Abstract

To standardize the tea export market and guarantee the interest of consumers, the adulteration problem in Taiping Houkui tea should be eliminated. In this study, a screening scheme comprising chemometrics and statistical analysis was proposed to estimate the geographical origin of Taiping Houkui tea. A total of 11 metal ions in Taiping Houkui tea were detected by performing a chemometric experiment. The key variables that can be used to identify the geographical origin of Taiping Houkui tea were screened using the least absolute shrinkage and selection operator method (LASSO). The statistical significance of selected key variables was also tested by analysis of variance (ANOVA), which confirmed the effectiveness of the LASSO. The proposed strategy was verified by the experimental testing and has great potential for determining the geographical origin of green tea.

Keywords:
tea quality estimation; feature extraction; least absolute shrinkage and selection operator (LASSO); analysis of variance (ANOVA)

Code	Origins	Specifications	Quantity	Plucking Time
#1	Houkeng village	50 g	20	2020-04-18
#2	Hougang village	50 g	20	2020-04-17
#3	Yanjia village	50 g	20	2020-04-17
#4	Sanhe village	50 g	20	2020-04-16
#5	Shihekeng village	50 g	20	2020-04-16
#6	Wangwangling village	50 g	20	2020-04-16

Reagents	Specification or grade	Source or manufacturer
HNO₃	Level-UP (68%)	Suzhou Jingrui Chemical Co., Ltd
H₂O₂	Degree of purity (30%)	Guoyao Group Chemical Reagent Co., Ltd.
Taiping Houkui tea standard materials	GBW10083	China Institute of Metrology
Pb, Cd, Cr, Cu, Fe, Mn, Zn, K, Na, Ga, Mg	1000 mg/L	National Center for Analysis and Testing of Nonferrous Metals and Electronic Materials
Sc, Ge, Im, Bi (Internal standard elements)	1000 mg/L	Ditto

Apparatus	Model	Country and Company
Inductively coupled plasma mass spectrometry	PE NeXlON 350D	Perkin Elmer Corporation, USA
High-pressure microwave digester	CEM MARS 5	CEM Corporation, USA
Ultra-pure water system	Millipore direct 16	Millipore Corporation, USA
Thermostatic drying chamber	M5	Binder Corporation, Germany
Temperature control electric heating plate	DB-3	Changzhou Guoyu Instrument Manufacturing Corporation, China
Ultrasonic water bath box	KQ-500DE	Kunshan Ultrasonic Instrument Corporation, China
Electronic analytical balance	Mettler-AL204-IC	Mettler Toledo Instruments Corporation, China

input: collect $X$ and $Y$ .
initialization: Let $\hat{β} = [β_{1}, β_{2}, \dots β_{m}] \in ℝ^{m \times 1}$ be an initial vector generated randomly, $l$ is the iterations.
1: for $d = 1,2, \dots D$
2: $l = 0$ ;
3: do { $l = l + 1$ ;
4: Cyclically For $j = 1,2, \dots m$
4: Update the $j_{t h}$ component ${\hat{β}}_{j}$ of $\hat{β}$ by ${\hat{β}}_{j} = \{\begin{cases} \frac{(ρ_{j} + \frac{λ_{d}}{2})}{z_{j}}, ρ_{j} < - \frac{λ_{d}}{2} \\ 0, - \frac{λ_{d}}{2} \leq ρ_{j} \leq \frac{λ_{d}}{2} \\ \frac{(ρ_{j} - \frac{λ_{d}}{2})}{z_{j}}, ρ_{j} > - \frac{λ_{d}}{2} \end{cases}$ , where
5: $ρ_{j} = \sum_{i = 1}^{n} (y_{i} - \sum_{c \neq j}^{m} {\hat{β}}_{c} x_{i, c}) x_{i, j}$ , $z_{j} = \sum_{i = 1}^{n} x_{i, j}^{2}$ ;
6: } while ( $‖β^{l - 1} - β^{l}‖ < 0.001$ \|\| $l$ < 500)
7: end for
8: Find the optimal $λ$ with the minimum $J (\hat{β})$ .
Output: The optimal coefficient vector $\hat{β} = \{β_{1}, β_{2}, \dots β_{m}\}$ corresponding to the optimal $λ$ .

Metal ions	Cd	Cr	Pb	Cu	Fe	Na	Zn	Ca	K	Mg	Mn
Cd	1.00
Cr	-0.07	1.00
Pb	0.89	-0.09	1.00
Cu	-0.36	0.17	-0.45	1.00
Fe	0.00	0.09	-0.05	0.08	1.00
Na	0.02	-0.29	0.17	-0.61	-0.04	1.00
Zn	-0.27	-0.28	-0.34	0.29	0.27	0.03	1.00
Ca	0.04	-0.05	-0.02	0.03	0.12	-0.26	0.04	1.00
K	0.10	0.01	-0.01	0.22	0.04	-0.12	0.22	0.37	1.00
Mg	0.23	-0.07	0.21	-0.24	0.08	0.32	0.17	0.36	0.79	1.00
Mn	-0.54	0.00	-0.67	0.50	-0.06	-0.27	0.21	0.33	0.40	0.04	1.00

Metal ions	Cd	Cr	Pb	Cu	Na
Coefficients	-2.31	-0.83	-4.94	0.47	0.04
p-values	2.12e-42	2.39e-7	6.68e-67	2.44e-41	3.00e-60

Method	Calibration set			Prediction set
Method	Median $R^{2}$	Median accuracy	Median RMSE	Median $R^{2}$	Median accuracy	Median RMSE
LASSO	0.882	0.679	0.577	0.842	0.611	0.656
PLS-DA	0.857	0.631	0.600	0.827	0.583	0.707

Sociedade Brasileira de Ciência e Tecnologia de Alimentos Av. Brasil, 2880, Caixa Postal 271, 13001-970 Campinas SP - Brazil, Tel.: +55 19 3241.5793, Tel./Fax.: +55 19 3241.0527 - Campinas - SP - Brazil
E-mail: revista@sbcta.org.br

Acompanhe os números deste periódico no seu leitor de RSS

[1] *Corresponding author: thpan@ahu.edu.cn