Unlocking the Aroma Profiles of Coffee Roasting Levels with an Electronic Nose Coupled with Machine Learning
Issued Date
2024-01-01
Resource Type
Scopus ID
2-s2.0-85201387765
Journal Title
Proceedings - 21st International Joint Conference on Computer Science and Software Engineering, JCSSE 2024
Start Page
678
End Page
681
Rights Holder(s)
SCOPUS
Bibliographic Citation
Proceedings - 21st International Joint Conference on Computer Science and Software Engineering, JCSSE 2024 (2024) , 678-681
Suggested Citation
Somaudon V., Kerdcharoen T. Unlocking the Aroma Profiles of Coffee Roasting Levels with an Electronic Nose Coupled with Machine Learning. Proceedings - 21st International Joint Conference on Computer Science and Software Engineering, JCSSE 2024 (2024) , 678-681. 681. doi:10.1109/JCSSE61278.2024.10613715 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/100599
Title
Unlocking the Aroma Profiles of Coffee Roasting Levels with an Electronic Nose Coupled with Machine Learning
Author(s)
Author's Affiliation
Corresponding Author(s)
Other Contributor(s)
Abstract
The dynamic shifts in the chemical composition of coffee with roasting have been successfully tracked using the electronic nose, representing its potential as a tool for profiling the aromatic complexity of coffee. Traditional methods have confirmed the physical transformation of beans during roasting, a well-known phenomenon. Complementing these findings, the e-nose demonstrates its efficacy by capturing the aromatic changes that occur throughout the roasting process. Furthermore, machine learning models applied to the e-nose data such as kNN, SVM, decision tree, and ANN, have shown promising results. Among these, the SVM model provides the greatest accurately reflecting the roasting profiles. This innovative, non-invasive approach provides a valuable alternative for the industry, paving the way for future applications in quality control and flavor profiling within the coffee industry.