Institutional Repository of Vilnius University
Klasikinių ir mašininio mokymosi metodų palyginimas kredito rizikos modeliavimui /
Abstract
dc:descriptionMachine learning methods are increasingly being used to address a variety of forecasting problems, including credit risk assessment. This paper examines how well 3 classification models – a classical logistic regression and two machine learning methods – decision trees and the XGBoost algorithm – can predict the probability of default for Swedish clients’ mortgage applications during the period from 2014 to 2018. The problem of data imbalance is taken into account: the models are estimated using several samples created by random undersampling. To avoid overfitting, k-fold cross-validation is used for the optimisation of machine learning algorithms, and the predictive power of the estimated models is compared using the area under the ROC curve. The results suggest that machine learning methods, such as XGBoost, may outperform traditional models in terms of more accurate predictions, but they pose a new challenge – a more difficult model interpretation.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2022
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Burkevičiūtė, Inesa,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD146237514&prefLang=en_US