Exploring Aroma Type and Alcohol Content Classification of Chinese Liquor Using Temperature-Modulated Electronic Nose

Chinese liquor, commonly referred to as baijiu, ranks among the world's most widely consumed and highest-selling spirits. Recent advancements in electronic nose (e-nose) technology have enabled the digital capture of odor profiles from various types of baijiu. However, distinguishing between aroma types remains challenging due to the intertwined nature of two critical attributes: aroma type and alcohol degree across many baijiu brands. In this study, we investigate e-nose signals from different baijiu aroma types across varying alcohol concentration levels using temperature-modulated metal oxide semiconductor gas sensors. By leveraging principal component analysis (PCA) for dimensionality reduction and support vector machines (SVM) for classification, we demonstrate that differentiating between aroma types is more intricate compared to distinguishing common alcohol concentration variations. This is evident from classification accuracy and silhouette scores. Our study underscores the need for developing an atlas of baijiu odor profiles to better understand aroma related features and enhance the specificity of e-nose-based brand classification. Our findings introduce a novel perspective on aroma type classification in baijiu, opening new avenues for practical applications in e-nose-based Chinese liquor classification models.