Predicting Sensory Quality of Long极ng Tea on the Basis of Physiochemical Data

文档简介Sensoryevaluationcanprovideintegrated,directmeasurementsoftheperceivedqualityoffoodproducts.However,asensorypanelissubjectiveandsuffersfrominconsistencyandinaccuracy.Inthispaper,weproposeasensoryevaluationsimulationmodelforLong极ngtea（aChinesebrandofgreentea）.ThephysiochemicalqualityindicatorsofLong极ngteaweredeterminedbyinstrumentalanalysis,includingcolor,aroma,andtaste.Meanwhile,thesensoryqualityoftheteawasevaluatedbyanexpertsensorypanel.Anartificialneuralnetworkwasconductedtoapproximatelypredictsensoryevaluationscoresonthebasisofphysiochemicaldata.Theresultsshowedthatphysiochemicalfactors,includinghue,fluorescencepeak5,huechromascale,b,L,3-（methylthio）propionaldehyde,α-terpineol,linalool,dimethylsulfide,totalaromavalue,caffeine,quinicacid,theanin,gallicacidandtotalcatechinswerebestcorrelatedwithsensoryevaluationscores.Furthermore,physiochemicalfeaturesthatwerechosenaccordingtoimportantfactorweightswereusedtoclassifyLong极ngteaintotwogrades.Experimentalresultsdemonstratedthatinstrumentalanalysiscouldbecomplementarilyusedintheevaluationandcontrolofsensoryqualitybyestablishingareasonablesensory-instrumentcorrelationandhuman-simulatedpredictivemodel.