Ghent University Academic Bibliography

Advanced

Prediction of specialty coffee cup quality based on near infrared spectra of green coffee beans

Kassaye Tolessa, Michael Rademaker, Bernard De Baets UGent and Pascal Boeckx UGent (2016) TALANTA. 150. p.367-374
abstract
The growing global demand for specialty coffee increases the need for improved coffee quality assessment methods. Green bean coffee quality analysis is usually carried out by physical (e.g. black beans, immature beans) and cup quality (e.g. acidity, flavour) evaluation. However, these evaluation methods are subjective, costly, time consuming, require sample preparation and may end up in poor grading systems. This calls for the development of a rapid, low-cost, reliable and reproducible analytical method to evaluate coffee quality attributes and eventually chemical compounds of interest (e.g. chiorogenic acid) in coffee beans. The aim of this study was to develop a model able to predict coffee cup quality based on NIR spectra of green coffee beans. NIR spectra of 86 samples of green Arabica beans of varying quality were analysed. Partial least squares (PLS) regression method was used to develop a model correlating spectral data to cupping score data (cup quality). The selected PLS model had a good predictive power for total specialty cup quality and its individual quality attributes (overall cup preference, acidity, body and aftertaste) showing a high correlation coefficient with r-values of 90, 90,78, 72 and 72, respectively, between measured and predicted cupping scores for 20 out of 86 samples. The corresponding root mean square error of prediction (RMSEP) was 1.04, 0.22, 0.27, 0.24 and 0.27 for total specialty cup quality, overall cup preference, acidity, body and aftertaste, respectively. The results obtained suggest that NIR spectra of green coffee beans are a promising tool for fast and accurate prediction of coffee quality and for classifying green coffee beans into different specialty grades. However, the model should be further tested for coffee samples from different regions in Ethiopia and test if one generic or region -specific model should be developed. (C) 2015 Elsevier B.V. All rights reserved.
Please use this url to cite or link to this publication:
author
organization
year
type
journalArticle
publication status
published
journal title
TALANTA
volume
150
pages
367 - 374
Web of Science type
J
Web of Science id
000370770500047
ISSN
0039-9140
DOI
10.1016/j.talanta.2015.12.039
UGent publication?
yes
classification
U
copyright statement
I don't know the status of the copyright for this publication
id
8520931
handle
http://hdl.handle.net/1854/LU-8520931
date created
2017-05-19 13:15:18
date last changed
2017-05-19 13:15:18
@article{8520931,
  abstract     = {The growing global demand for specialty coffee increases the need for improved coffee quality assessment methods. Green bean coffee quality analysis is usually carried out by physical (e.g. black beans, immature beans) and cup quality (e.g. acidity, flavour) evaluation. However, these evaluation methods are subjective, costly, time consuming, require sample preparation and may end up in poor grading systems. This calls for the development of a rapid, low-cost, reliable and reproducible analytical method to evaluate coffee quality attributes and eventually chemical compounds of interest (e.g. chiorogenic acid) in coffee beans. The aim of this study was to develop a model able to predict coffee cup quality based on NIR spectra of green coffee beans. NIR spectra of 86 samples of green Arabica beans of varying quality were analysed. Partial least squares (PLS) regression method was used to develop a model correlating spectral data to cupping score data (cup quality). The selected PLS model had a good predictive power for total specialty cup quality and its individual quality attributes (overall cup preference, acidity, body and aftertaste) showing a high correlation coefficient with r-values of 90, 90,78, 72 and 72, respectively, between measured and predicted cupping scores for 20 out of 86 samples. The corresponding root mean square error of prediction (RMSEP) was 1.04, 0.22, 0.27, 0.24 and 0.27 for total specialty cup quality, overall cup preference, acidity, body and aftertaste, respectively. The results obtained suggest that NIR spectra of green coffee beans are a promising tool for fast and accurate prediction of coffee quality and for classifying green coffee beans into different specialty grades. However, the model should be further tested for coffee samples from different regions in Ethiopia and test if one generic or region -specific model should be developed. (C) 2015 Elsevier B.V. All rights reserved.},
  author       = {Tolessa, Kassaye and Rademaker, Michael and De Baets, Bernard and Boeckx, Pascal},
  issn         = {0039-9140},
  journal      = {TALANTA},
  pages        = {367--374},
  title        = {Prediction of specialty coffee cup quality based on near infrared spectra of green coffee beans},
  url          = {http://dx.doi.org/10.1016/j.talanta.2015.12.039},
  volume       = {150},
  year         = {2016},
}

Chicago
Tolessa, Kassaye, Michael Rademaker, Bernard De Baets, and Pascal Boeckx. 2016. “Prediction of Specialty Coffee Cup Quality Based on Near Infrared Spectra of Green Coffee Beans.” Talanta 150: 367–374.
APA
Tolessa, K., Rademaker, M., De Baets, B., & Boeckx, P. (2016). Prediction of specialty coffee cup quality based on near infrared spectra of green coffee beans. TALANTA, 150, 367–374.
Vancouver
1.
Tolessa K, Rademaker M, De Baets B, Boeckx P. Prediction of specialty coffee cup quality based on near infrared spectra of green coffee beans. TALANTA. 2016;150:367–74.
MLA
Tolessa, Kassaye, Michael Rademaker, Bernard De Baets, et al. “Prediction of Specialty Coffee Cup Quality Based on Near Infrared Spectra of Green Coffee Beans.” TALANTA 150 (2016): 367–374. Print.