Detection of adulteration in camellia seed oilses

文档简介Anelectronicnosewasusedforthedetectionofmaizeoiladulterationincamelliaseedoilandsesameoil.Theresultsofmultivariateanalysisofvarianceshowedthatthesensorsignalsofdifferentkindsofoilaresignificantlydifferentfromeachother.Prin-cipalcomponentanalysisPCAcannotbeusedtodiscriminatetheadulterationofcamelliaseedoil,butcanbeusedinthediscriminationofadulterationinsesameoil.LineardiscriminantanalysisLDAismoreeffectivethanPCAandcanbeusedinadulterationdiscriminationforbothcamelliaseedoilandsesameoil.InordertocheckthediscriminativepowerofLDA,canonicaldiscriminantanalysiswasperformedaswell.Acceptableresultswerealsoobtained:Theaccuracyofpredictionwas83.6%forcamelliaseedoiland94.5%forsesameoil.TheartificialneuralnetworkANNmodelwasusedtodetectthepercentageofadulterationincamelliaseedoilandsesameoil.Theresultsshowedthat,basedonANNasitspatternrecognitiontechnique,theelec-tronicnosecannotpredictthepercentageofadulterationincamelliaseedoil,butcanbeusedinthequantitativedeterminationofadulterationinsesameoil.