Analysis of main aroma components of Zherong white tea and establishment of machine learning model for sensory evaluation

Abstract Background Zherong white tea, a distinctive specialty from Fujian Province, has attracted significant consumer attention due to its unique aroma and flavor profile. In this study, we utilized advanced analytical techniques and chemometric methods to conduct a thorough analysis of the volatile aroma compounds in Zherong white tea. Results Using proton transfer reaction‐time‐of‐flight mass spectrometry ( PTR ‐ TOF ‐ MS ), we identified a diverse array of aroma compounds, including alcohols, aldehydes, ketones, and esters, as the primary contributors to its aroma. Further chemometric analyses, such as principal component analysis ( PCA ) and orthogonal partial least squares discriminant analysis ( OPLS ‐ DA ), revealed key aroma compounds, such as linalool and dimethyl sulfide, which effectively distinguish Zherong white tea from other white tea varieties. Additionally, a backpropagation ( BP ) predictive model based on machine learning algorithms was developed to accurately predict sensory aroma scores, providing a more objective and efficient alternative to traditional evaluation methods. This model significantly improves the precision and efficiency of sensory assessment, addressing the inherent subjectivity and limitations of conventional tea quality evaluation. Conclusion In summary, PTR ‐ TOF ‐ MS combined with PCA and OPLS ‐ DA , among other analytical techniques, enables more effective identification of the 18 key aroma compounds in Zhongrong white tea. The BP model demonstrates high accuracy in predicting aroma sensory scores, as evidenced by a coefficient of determination ( R 2 ) value approaching 1. © 2025 Society of Chemical Industry.