電子鼻識別中的預處理與模糊神經網絡算法.pdf

1、大連理工大學碩士學位論文電子鼻識別中的預處理與模糊神經網絡算法姓名：楊俊申請學位級別：碩士專業(yè)：物理電子學指導教師：閆衛(wèi)平20071220大連理1=大學碩十學位論文ArithmeticofSignalPreprocessingandFuzzyNeuralNetworkinRecognitionofElectronicNoseAbstractCausedbythephysicalshortcomingsofgassL=Tlsors，iti

2、simpossibleforasinglegassL_flSOrtoidentifymultiplegasesSotheelectronicnosetechniquesbasedongassensorarrayandpatternrecognitionisbecominganimportantwayindealingwithcross—sensitivityingasanalysisThepatternrecognitiontechno

3、logyplaysthecrucialroletothebehaviorofelectronicnoseAimedtothisrequirement，theresearchstudiesseveralrelatedpatternrecognitiontechniquesinordertealizethequantitativelyrecognitionofthemixedgasesTheresearchstudiesthePrincip

4、alComponentsAnalysisandIndependentComponentAnalysisarithmeticTheprincipalcomponentanalysiscouldsubstituteasmallnumbercomprehensivevariablefororiginalmultidimensionalvariableandobviouslysimplifydatastructureinthepre—condi

5、tionofminimizinglossoforiginaldatainformationTheresearchcombinesthePrincipalComponentsAnalysisandIndependentComponentAnalysisinordertealizeaneffectivedatapre—processingintheelectronicnosesystemSomeexperimentsareperformed

6、toverifytheoptimizingeffectofPCAandICAAimedtotheneuralnetworkandthefuzzy109icofthepatternrecognitiontechnique，combinestheIogicalexpressionabilityofthefuzzysystemandtheselfstudingabilityoftheneuralnetwork，bringsforwardafu

7、zzyneuralnetworkarithmetic：Takagi—SugenofuzzyneuralnetworktorealizethequantitativelyrecognitionofthealcoholthicknessBasedonthismethod，bringsforwardanadvancedTakagi—SugenoFuzzyLogicSystembasedonNeuralNetworksitusestheneur

8、alnetworktosimulatethesubjectiveselectionThismethodCanavoidthelowefficiencyandimpreciseofthesubjectiveselectionandobtaintheassociatedmembershipfunctionandreasoningregulationaccuratelyTheresearchusesthealcoholelectronicno

9、setestingdateastheswatchusingthearithmeticmentionedabovetosimulateandtest54kindsofmixedgasestheaccuraterateofthequantitativelyrecognitionCanachievetoabove99％111eresultprovesthatthesignalpreprocessingtechniquesandthefuzzy