服務(wù)熱線：010-83993593手機號碼：13810615661
地 址：北京市西城區(qū)廣安門外大街168號朗琴國際大廈B座603室
| Electronic nosedata analysis for detection of mai | ||||||||||||
|
||||||||||||
| an!°electronicnose!±hasbeenusedforthedetectionofadulterationsofsesameoil.thesystem,comprising10metaloxidesemiconductsensors,wasusedtogenerateapatternofthevolatilecompoundspresentinthesamples.priortodifferentsupervisedpatternrecognitiontreatments,featureextractiontechniqueswereemployedtochooseasetofoptimaldiscriminantvariables.principalcomponentanalysispca,fisherlineartransformationflt,stepwiselineardiscriminantanalysisstep-lda,se-lectionbyfisherweightssfwwereused,respectively.andthen,lineardiscriminantanalysislda,probabilisticneuralnetworkspnn,backpropagationneuralnetworksbpnnandgeneralregressionneuralnetworkgrnnwereappliedaspatternrecognitiontechniquesfortheelectronicnose.asforldaandpnn,fltwasthemosteffectivefeatureextractionmethod,whilestep-ldawasthemosteffectivewayforbpnnandfltwasmoresuitableforgrnn.withonlyonesamplemisclassi?edinourexperiment,ldaismorepowerfulthanpnn.excellentresultswereobtainedinthepredictionofpercentageofadulterationinsesameoilbybpnnandgrnn.aftertrainingforsometime,bpnncouldpredicttheadulterationquantitativelymorepreciselythangrnn,whereaswithfltasitsfeatureextractionmethodandwithoutiterativetraining,grnncouldalsoyieldratheracceptableresults.?2006elsevierb.v.allrightsreserved. |
||||||||||||
