北京盈盛恒泰科技有限責(zé)任公司是一家專業(yè)從事食品分析檢測儀器和環(huán)境應(yīng)急檢測儀器的銷售技術(shù)支持和售后服務(wù)工作公司目前是日本INSENT日本JWP意大利VELP美國FTC德國AIRSENSE德國OWR美國UNITY美國Navas美國FiltationEngineering等世界儀器制造商的ZG區(qū)營銷合作伙伴和技術(shù)服務(wù)ZX，產(chǎn)品覆蓋食品感官分析食品營養(yǎng)分析食品安全檢測及環(huán)境應(yīng)急監(jiān)測等公司在香港北京上海廣州成都海口銀川等地設(shè)有分公司和辦事機(jī) Aninvestigationwasmadetoevaluatethecapacityofanelectronicnose(E-nose,PEN2)toclassifytheteaqualitygrade.Fourteagroups(A120,A280,A380andA600)withadifferentqualitygradewereemployed.Intheexperiment,thevolumeofvialandtheheadspacegeneratedtimewereconsideredcorrespondingtothe5gteasamples,andtheoptimumexperimentalprocedurewasdeterminedbyusingthevarianceanalysisANOVA),multivarianceanalysisandprinciplecomponentsanalysis(PCA).Theresultsshowedthatthevolumeofvialaffectedtheresultofdiscrimination,andtheheadspacegeneratedtimehadnosigni?canteffectontheE-noseresponse.ThefourteagroupsweremeasuredandresponsevaluesatfourdifferentcollectiontimeswereconductedbyPCA,lineardiscriminationanalysis(LDA)andarti?cialneuralnetwork(ANN).OnlyA120,A380andA600couldbediscriminatedbyPCA.However,thefourteagroupswerediscriminatedcompletelybyLDA.TheresponsevalueoftheE-noseat60swasoptimumtobeusedfordiscrimination.ThemethodofANN(networktopology20-12-4)wasperformedand90%oftheotalteasampleswereclassi?edcorrectlybyusingtheback-propagationneuralnetwork.?2006ElsevierB.V.Allrightsreserved. 電子鼻-Z傳感器生產(chǎn)商產(chǎn)品