Kliento nemokumo rizikos lygio vertinimas ir prognozavimas /

Assessing and predicting the level of insolvency risk is an essential process in the financial world, providing businesses with necessary insight into potential risks and their possible consequences. Modern economic conditions and a changing market demand constant improvement of the risk management and predicting process. The main objective of this study is to develop a classifier to assess and predict the level of a customer’s default risk. In the theoretical part of this research, literature was examined to find information on methods and classifiers applied to assess and predict insolvency risk level. The practical part involves primary data analysis, cluster analysis, and construction and evaluation of machine learning classifiers. Primary data analysis identified key factors influencing the customer insolvency risk level. During cluster analysis, 3 clusters were obtained with different financial indicator values. The first cluster consisted of companies with the lowest risk level, the second - medium, and the third - high. By applying various machine learning models and selecting important features in different ways, the highest results were achieved using a logistic regression classifier with cluster analysis data. The accuracy of this classifier is 0,87, precision is 0,89, recall, F1 score is 0,87, MCC is 0,7, AUC > 0, 8.

Grantor dc:publisher

Institutional Repository of Vilnius University

Year dc:date

2024

Author dc:creator

• Valeikaitė, Austėja,

Statement dc:rights

• info:eu-repo/semantics/openAccess

Language dc:language

lit

Repository record dc:identifier

https://repository.vu.lt/VU:ELABAETD210643738&prefLang=en_US

OAI identifier oai:identifier

oai:vu.lt:elaba:210643738