| dc.contributor.author | Akgün Dehiller, Dennis | |
| dc.contributor.author | Arif, Milan | |
| dc.date.accessioned | 2022-06-29T09:19:26Z | |
| dc.date.available | 2022-06-29T09:19:26Z | |
| dc.date.issued | 2022-06-29 | |
| dc.identifier.uri | https://hdl.handle.net/2077/72392 | |
| dc.description | MSc in Finance | en_US |
| dc.description.abstract | The purposeofthispaperistoexaminewhetherwecanpredictbankruptciesinSwedish
industrial andmanufacturingSMEsbeforetheyoccur.Toconductthepredictionsweuse
a traditionalstatisticalmodel,multiplediscriminantanalysis(MDA)andamoremodern
machinelearningapproachinwhichweutilizesupportvectormachines(SVM).Wethen
compare thetwomodelsinordertofindoutwhichofthemperformsthebest.Further,we
examine howmanyyearspriortofailureweareabletopredictthebankruptcy,andthus
test bothmodelsuptofiveyearspriortofailure.Ourresultssuggestthatweareableto
predict bankruptcieswithsignificantlyhigheraccuracythana50/50guessingstrategy.
Further,wefoundwewereabletopredictbankruptciesupthefiveyearspriortothem
occurring,usingboththetraditionalMDAmodelandtheSVMmodel.Whencomparing
the twowefoundtheMDAmodelachievehigherpredictionaccuracyoneyearpriorto
failure. | en_US |
| dc.language.iso | eng | en_US |
| dc.relation.ispartofseries | 2022:152 | en_US |
| dc.title | Predicting bankruptcies in Swedish manufacturing firms A comparison between traditional statistical and machine learning methods | en_US |
| dc.type | Text | |
| dc.setspec.uppsok | SocialBehaviourLaw | |
| dc.type.uppsok | H2 | |
| dc.contributor.department | University of Gothenburg/Graduate School | eng |
| dc.contributor.department | Göteborgs universitet/Graduate School | swe |
| dc.type.degree | Master 2-years | |
