Development of regression model to differentiate quality of black tea (Dianhong): correlate aroma properties with instrumental data using multiple linear regression analysis

Abstract

SummaryTo develop a flexible and objective method to discriminate Dianhong tea quality level, the instrumental analysis of odour‐active compounds of three commercial grades of Dianhong tea and quantitative descriptive analysis of their overall aromas were conducted. Eight aroma properties were then statistically correlated with odorants using stepwise multiple linear regression analysis. Thirty‐nine odour‐active components were screened in Dianhong tea infusion, among which dimethyl disulphide and 1‐(1H‐pyrrol‐2yl)‐ethanone were identified for the first time. Adjusted R2 of multiple regression models for eight sensory properties were all >0.95, except ‘resinous’ attribute (R2 = 0.834), and RMSE were <0.18 for all. Regression models showed that 2‐methylbutanal, 3‐methylbutanal, linalool, geraniol, linalool oxides, phenyl acetaldehyde, methyl salicylate, ethyl 2‐methylbutyrate, 3,7‐dimethyl‐3‐octanol, benzaldehyde and 6,10‐dimethyl‐5,9‐undecadien‐2‐one positively contributed to the Dianhong tea aroma profile, while 1‐penten‐3‐ol, (Z)‐3‐hexen‐1‐ol and (E,E)‐2,4‐heptadienal were negative aroma contributors. Regression model development for Dianhong tea aroma quality would provide a uniform standard for objectively assessing tea quality among different laboratories and cultures.