Related Experiment Video
Updated: May 5, 2026
13:02
The Terroir Concept Interpreted through Grape Berry Metabolomics and Transcriptomics
Published on: October 5, 2016
9.7K
Machine Learning and SHAP Feature Analysis: Classification Model for Aroma Components in Green Plum Wine
Xuhui Zhang1, Mengsheng Deng1, Yu Lei1
1School of Food and Liquor Engineering, Sichuan University of Science and Engineering, Yibin 644000, China.
Foods (Basel, Switzerland)
|May 4, 2026
Summary
This study used machine learning to analyze fermented green plum wine flavors. Ethyl octanoate and other esters significantly contribute to the distinct floral and fruity aromas identified.
Area of Science:
- * Food Science and Technology
- * Analytical Chemistry
- * Machine Learning Applications
Background:
- * Fermented fruit wines, like green plum wine, possess complex volatile flavor profiles.
- * Understanding these profiles is crucial for quality control and product development.
- * Traditional methods for flavor analysis can be time-consuming and lack comprehensive interpretation.
Purpose of the Study:
- * To systematically investigate and differentiate volatile flavor profiles in fermented green plum wines.
- * To evaluate the effectiveness of machine learning (ML) algorithms for flavor profiling and classification.
- * To identify key volatile compounds contributing to the distinct aromas using SHapley Additive exPlanations (SHAP).
Main Methods:
- * Integration of Gas Chromatography-Mass Spectrometry (GC-MS) for volatile compound identification.
- * Sensory evaluation and Odor Activity Value (OAV) analysis to assess aroma impact.
- * Application of machine learning algorithms (including fuzzy c-means clustering and decision tree models) and SHAP for data interpretation.
Main Results:
- * Floral and fruity aromas were predominant, with esters like ethyl benzoate and ethyl octanoate being major contributors.
- * Fuzzy c-means clustering successfully categorized wines into three distinct flavor groups.
- * The decision tree model achieved high accuracy (95.13%) in flavor classification, with ethyl octanoate, benzyl ethanoate, and 2-phenylethyl ethanoate identified as key influential compounds by SHAP analysis.
Conclusions:
- * Machine learning provides a powerful and efficient approach for classifying and interpreting complex flavor profiles in fermented beverages.
- * The study demonstrates the successful application of ML and SHAP in identifying key aroma contributors in green plum wine.
- * These findings have broad implications for flavor science and the quality assessment of fermented fruit products.
