Analysis of aroma characteristics of the aged Shaoxing Huangjiu by a combination of multiple instruments and machine learning

Aging governs the aroma quality and economic value of Huangjiu. The traditional aging process can easily result in the common fraudulent practice that the young wine was labeled as the aged wine. In this study, the effect of aging time on the aroma characteristics of Shaoxing Huangjiu with different aging years (3, 5, 8, and 10) were investigated using a combination of multiple instruments and machine learning (ML). With aging times prolonged, free amino acids contents presenting the bitter, sweet, and umami increased. The alcohol content became lower, whereas those of aldehydes, esters, and ketones increased. By ML, the overall information of volatiles was maximized and analyzed systematically. The performance of ML models was evaluated by the confusion matrix. Results revealed that key volatile aroma compounds (e.g. 2-phenylethyl alcohol, nonanal, and isoamyl alcohol) dominated the overall aroma characteristics, and were viewed as reliable markers for distinguishing different aged Shaoxing Huangjiu . • Aroma fingerprints of aged Shaoxing Huangjiu were evaluated by GC-IMS. • Aging process could affect the overall aroma characteristics of Shaoxing Huangjiu. • Basic FAA contributed to the flavor formation of Shaoxing Huangjiu during aging. • Dominant volatile aroma compounds were identified by OAV analysis • By machine learning, the differential markers of Shaoxing Huangjiu were screened.