Construction of quadratic multinomial regression model for Guizhou green tea taste attributes

【Objective】To analyze the sensory taste characteristics,the content of main taste components and their Dot values of Guizhou green tea,and to explore the intelligent sensory taste characteristics and related chemical substances,and clarify the chemical contribution mechanism of the formation of the taste characteristics of Guizhou green tea,so as to provide theoretical reference for establishing a digital evaluation of Guizhou green tea.【Method】Based on 39 green tea samples from 5 main tea producing areas in Guizhou(Zunyi,Tongren,Guiyang,Qiannan,Liupanshui),artificial sensory evaluation,intelligent sensory evaluation,chemical composition detection were conducted,combined with quadratic multinomial regression model analysis method was used to construct the evaluation model of Guizhou green tea taste attributes.【Result】Mellowness and freshness were typical common attributes of Guizhou green tea.The intelligent sensory perception of Guizhou green tea from different producing areas had high responsivity values for umami and astringency,which were 12.85-18.21 and 13.15-26.77,respectively;bitterness responsivity were lower,ranging from 3.98 to 8.44.Guizhou green tea from different producing areas had little difference in the content of taste chemical components,among which the average mass fractions of water extract,total free amino acids and tea polyphenols were 45.15%-46.63%,5.49%-6.46%and 25.56%-27.63%,respectively.The average contents of caffeine and chlorogenic acid were 35.85-41.63 mg/g and 1.96-2.53 mg/g,respectively.An evaluation model for the taste attributes of Guizhou green tea was constructed:Y(bitter)=-0.3041+0.4189X₆+0.2182X₂²+0.0015X₁X₃-0.0016X₁X₄-0.0885X₂X₆+0.6868X₂X₇-0.0434X₆X₇;Y(astringent)=21.9559-0.4562X₆+0.2728X₂²-0.0196X₁X₆+0.1206X₁X₇+0.0192X₃X₆-0.0308X₄X₇;Y(fresh)=13.3301+0.1096X₂²-0.0037X₁X₆+0.0058X₃X₆-0.0013X₄X₆(X₁,X₂,X₃,X₄,X₆and X₇respectively represented theanine,glutamic acid,caffeine,EGCG,GCG,chlorogenic acid).The quadratic polynomial regression model showed that the correlation coefficient was above 0.8100,the coefficient of determination was above 0.6600,and the taste intensity of the reference model was 9.62-23.10.The bitterness was 0.69-10.99,and the astringency was 2.95-40.51.【Conclusion】High water extract,high amino acid,high caffeine and moderate tea polyphenols form the thickness,mellow and refreshing taste characteristics of Guizhou green tea.According to the three models,the highest intensity and the lowest intensity of the taste attributes of Guizhou green tea can be combined.