基于LVQ神经网络的财务舞弊识别模型研究

《基于LVQ神经网络的财务舞弊识别模型研究》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

1、东南人学硕上学位论文THERESEARCHoNTHEMoDELoFDETECTINGTHEFRAUDULENTFINANCIALSTATEMENTSBASEDONLVQNEURALNETWORKGraduate：GuNing-shengSupervisor：FengQin-ehaoSoutheastUniversityAbstractFinancialreportingfrauddoesnotonlyaffecttheinterestsofinvestors，butalsothreatthecapitalm

2、arketandeconomy．Governmentsofallcountrieshadtakenvariousmeasurestocurbthefraudulentfinancialstatements，butthephenomenonoffraudulentfinancialstatementsdidn’tdisappear．Fraudulentfinancialstatementsarestillthreateningtheinvestorsandthecapitalmarket．Soinordert

3、oprotecttheinvestorsandregulatethedomesticcapitalmarkets，thestudyoffinancialfrauddetectionisveryimportant．Inordertofindthevariablesofdetectingthefraudulentfinancialstatements，themotiveandmeansoffinancialfraudhasbeenanalyzed．Corporategovernanceandownerships

4、tructurealsohasbeenanalyzed．Bytheanalysis，66alternativeindicatorsofdetectingfinancialfraudhavebeenfound．InordertOmakesurewhichindicatorcandetectthefraudulentfinancialstatements，anempiricalresearchhasbeendone．Throughthisempiricalresearch,31indicatorswhichha

5、vetheabilityofdetectingthefraudulentfinancialstatementshavebeenfound．ThentheprincipalcomponentmethodWasusedtoreducethenumberofindicators．Throughit，12comprehensiveindicatorswereacquired．Thenthrough60modelingsamplesand32testsamples，themodelofdetectingthefrau

6、dulentfinancialstatementswasbuiltbytheLVQnetwork．ThenaempiricalresearchabouttestfI，earningVectorQuantization)neuralsampleshavebeendone，theaccuracyofthemodelbuiltbyLVQneuralnetworkWast090．6％．Atlastthreeothermodelsofthefraudulentfinancialstatementsupdetectin

7、gwerebuiltbytheBPneuralnetwork,GA—MLPneuralnetworkandSVM．Throughcomparingthefourmodels，thispaperfoundthatthemodelbuiltbyLVQneuralnetworkhasthehighestpredictionprecisionaboutthetestsamples．Keywords：fraudulentfinancialstatements；detection；LVQneuralnetwork；su

8、pportvectormachines(SVM)II目录摘要⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯．IAbstract．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．．II第一章导论⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯l1．1研究背景⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯⋯．11．2研究财务舞弊识别的意义⋯⋯⋯