Predicering av låntagares återbetalningsförmåga med hjälp av maskininlärningsmetoder : En jämförelse av metoderna logistisk regression, random forest, K-nearest neighbor och support vector machines
| Autor: | Leth, Jakob, Ahlberg, Ellen |
|---|---|
| Jazyk: | švédština |
| Rok vydání: | 2020 |
| Předmět: | |
| Popis: | This thesis aims to investigate how statistical machine learning methods can be used to predict an individual's risk of default with regards to chosen model evaluation parameters. Logistic regression, random forest, K-nearest neighbor and support vector machines were the investigated techniques. The methods were applied on a dataset from the international consumer finance provider Home Credit Group. The results show that none of the implemented models give useful predictions for customers default risk. The reason is that all models struggle to identify individuals who do not repay their loans. The thesis concludes that an improved variable selection method and enhanced data processing probably could increase the accuracy of the models. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|