Đang tải... (xem toàn văn)
MINISTRYOFEDUCATION&TRAINING STATEBANKOFVIETNAM HOCHI MINHCITYUNIVERSITYOFBANKING NGUYENTHINGOCANH APPLICATION OF MACHINE LEARNING FORPREDICTINGPROBABILITYOFDEFAULTOFS MALLANDMEDIUMENTERPRISES GRADUAT[.]
MINISTRYOFEDUCATION&TRAINING STATEBANKOFVIETNAM HOCHI MINHCITYUNIVERSITYOFBANKING -NGUYENTHINGOCANH APPLICATION OF MACHINE LEARNING FORPREDICTINGPROBABILITYOFDEFAULTOFS MALLANDMEDIUMENTERPRISES GRADUATION THESISMAJOR: FINANCE – BANKINGCODE:7340201 HOCHIMINHCITY,2022 MINISTRYOFEDUCATION&TRAINING STATEBANKOFVIETNAM HOCHI MINHCITYUNIVERSITYOFBANKING NGUYENTHINGOCANH APPLICATION OF MACHINE LEARNING FORPREDICTINGPROBABILITYOFDEFAULTOFS MALLANDMEDIUMENTERPRISES GRADUATION THESISMAJOR: FINANCE – BANKINGCODE:7340201 SUPERVISOR Ph.D NGUYEN MINH NHATHOCHIMINHCITY,20 22 ABSTRACT Corporate default predictions play an essential role in each sector of the economy, ashighlighted by the Covid - 19 pandemic The recent high incidence ofSmall andMedium Enterprises bankruptcies has highlighted the necessity of anticipating defaultsin many sectors Based on the importance and necessity, this study aims to investigatewhatappropriatemodelsforpredictingtheprobability ofdefaultofSMEsintheVietnameseCommercialBanksSystembyMachineLearningapproache s;howtochoose an appropriate model for predicting the probability of default of SMEs in theVietnamese Commercial Banks System by Machine Learning approaches; and how tochoose an appropriate model for predicting the probability of default of SMEs in theVietnameseCommercialBanksSystembyMachineLearningapproachesusingaunique database of 400 Vietnamese SMEs over the 2019 – 2021 period including 13independentf i n a n c i a l v a r i a b l e s T h e m o s t s i g n i f i c a n t c o n t r i b u t i o n o f t h i s r e s e a r c h i s theapplicationofMachineLearningapproachesintheuseoffinancialindicatorstoanticipate the default likelihood of SMEs, as a result, leading to improved efficiencyoutcomesincommercialbanks'creditriskcontrolinVietnaminthefuture.Thisresear ch analyzes theperformanceofa setofMachine Learning(ML)models inpredicting default risk, using a standard statistical model, in particular, the LogisticRegression Model When just a restricted amount of information is provided, such as inthecaseoffinancialindicators,MLmodels(DecisionTreeandRandomForest)outperforms tatisticalmodelswasfoundintermsofdiscriminatorypowera n d precision matrix and F1 – Score are used to evaluate themostappropriatetopredicttheprobabilityofdefaultofSMEs which Confusion model is DECLARATION I declare that this thesis has been composed solely by myself and that it has not beensubmitted, in whole or in part, in any previous application for a degree Except wherestates otherwise by reference or acknowledgment, the work presented is entirely myown Theauthor NGUYEN THINGOCANH ACKNOWLEDGEMENTS Throughout the writing of this thesis, I have received a great deal of support andassistance Firstandforemost,Iwouldliketoexpressmyheartfeltgratitudeandp r o f o u n d gratitude to the professors of the Ho Chi Minh City University of Banking for theirpassionate teaching and for solidifying the firm foundation of knowledge that enabledmetosuccessfullyfinishtheuniversityprogram Second, I would like to acknowledge and give my warmest thanks to my supervisor,Ph.D Nguyen Minh Nhat for providing me with thorough instruction and unwaveringsupport in finishing my graduation thesis It would be tough for me to accomplish mythesiswithouthiscarefulassistance Because of my limited practical experience, the topic of my graduation thesis cannotavoid some faults; nonetheless, I am looking forward to obtaining more advice fromlecturers to gain new experiences These experiences, I feel, will be highly beneficial tomyfuturedevelopment I sincerely thank you! NguyenThiNgocAnh TABLEOFCONTENT LISTOFABBREVIATIONS vii LISTOF FIGURES viii LISTOFTABLES .ix CHAPTER1INTRODUCTION 1.1 Theurgency oftheresearch .1 1.2 ResearchObjectives 1.3 ResearchQuestions 1.4 ResearchSubjectandScope .6 1.4.1 ResearchSubject 1.4.2 ResearchScope 1.5 ResearchContributions 1.6 ResearchMethodology 1.7 TheStructure ofResearch CHAPTER2LITERATUREREVIEW 10 2.1 ProbabilityofDefault(PD) .10 2.2 Overviewofthemodelsusedtopredictthe ProbabilityofDefault ofSMEs 11 2.2.1 TheStructuralModels 11 2.2.1.1 RegressionAnalysisModels .12 2.2.1.2 DiscriminantAnalysisModels 12 2.2.1.3 LogisticModels 13 2.2.2 TheNon-StructuralModels 14 2.2.2.1 Decision Tree Model(DT) 14 2.2.2.2 RandomForestModel(RF) 15 2.2.2.3 Artificial Neural Network Models (ANNs) 15 2.2.2.4 EnsembleLearning 16 2.3 Previous RelatedResearch .17 CHAPTER3 DATA ANDMETHODOLOGY 20 3.1 MethodologicalModelFramework 20 3.2 Data collection 21 3.3 InputVariablesSelection 22 3.4 TheProbabilityof Defaultprediction models 25 3.4.1 LogisticRegressionModel 25 3.4.2 Decision TreeModel(DT) 26 3.4.3 RandomForestModel(RF) 28 3.4.4 Confusion Matrix 29 3.4.5 F1-Score 31 CHAPTER4 EMPIRICALRESULTS 32 4.1 Descriptivestatisticsresults .32 4.2 Correlations 33 4.3 Regressionresultsofaparametric model 34 4.3.1 LogisticRegressionResult 34 4.3.2 Confusionmatrixoftheparametricmodel .37 4.4 Regressionresultsofnon-parametricmodels 38 4.4.1 DecisionTree .38 4.4.2 RandomForest 40 4.5 RegressionresultofEnsembleLearning 41 CHAPTER5CONCLUSIONANDRECOMMENDATION .43 5.1 Applying themodeltoforecastthelikelihoodofdefaultforSMEcustomersatVietnameseCo mmercialbanks 43 5.1.1 Toolst h a t a i d i n t h e i d e n t i f i c a t i o n o f g r o u p s o f p r o s p e c t i v e S M E s customers .43 5.1.2 The model resultsserveasthefoundationforcreditpolicyorientation 44 5.1.3 Applyingt h e modelr e s u l t s t o i m p r o v e c r e d i t r i s k m a n a g e m e n t efficiencyinCommercialbanks .45 5.2 Applyingth emodelto anticipatethelikelihood ofdefault for Cre ditRatingAgencies inVietnam 46 5.3 Topiclimitationandpotentialresearchdirections 48 5.3.1 Topiclimitation 48 5.3.2 Potentialresearchdirections 49 REFERENCES i APPENDIX .vii LISTOFABBREVIATIONS Number Symbol English PD ProbabilityofDefault SMEs Smalland MediumEnterprises VCCI VietnamChamberofCommerceandIndustry ANN ArtificialNeuralNetworks RF RandomForestModel DT DecisionTreeModel MDA MultivariateDiscriminantAnalysis AAFS AnnualAuditedFinancialStatements LISTOFFIGURES Figure1.1:Forecast InsolvencyGrowthin2022comparedto2019 Figure3.1:Aproposedschematicprocessofthemethodologyofthisstudy 20 Figure3.2:Simulationofthedecisiontreemodel .27 Figure3.3:RandomForestSimplified 28 Figure4.1:CorrelationMatrix .34 Figure4.2:RegressionresultoftheDecisionTreemodel 38 Figures1:SixstepswiththecodestorunLogisticRegressionModel xi Figure2:TheusagecodetodrawDecisionTreemodel .xii Figure3:TheusagecodetoruntheConfusionMatrixofDecisionTreemodelxiiFigure4:Th eusagecodetorunConfusionMatrixofRandomForestmodel xii Figure5 : T h e u s a g e c o d e t o r u n t h e C o n f u s i o n M a t r i x o f E n s e m b l e L e a r n i n g model xii