Napovedovanje predčasnega odplačevanja kreditov

BENČINA, MIHA (author), Kononenko, Igor (mentor), Cvetković, Božidara (co-mentor)

Ena od najbolj razširjenih dejavnosti bank je dajanje kreditov. Največje tveganje je, da dolžnik kredita ne bo sposoben odplačati. Manjši problem za banko se pojavi, ko dolžnik kredit odplača predčasno. V diplomskem delu smo se osredotočili na slednji problem in poizkušali s čim večjo uspešnostjo napovedati, kateri krediti bodo v naslednjem kvartalu predčasno odplačani. Na podlagi bančnih in eksternih makroekonomskih podatkov smo konstruirali napovedne modele, ki temeljijo na različnih metodah. Rezultati razkrijejo, da smo pri zastavljenih omejitvah, uspešno odkrili približno polovico predčasno odplačanih kreditov.

• Language: Slovenian
• Keywords: kredit, predčasno odplačevanje, napovedni model, klasifikacija
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: FRI - Faculty of Computer and Information Science
• Year: 2020
• PID: 20.500.12556/RUL-117190
• COBISS.SI-ID: 21607683
• Publication date in RUL: 01.07.2020

Secondary language

• Language: English
• Title: Predicting early loan repayments

Abstract

One of the most common fields of banking is lending money in the form of a loan. The biggest risk regarding that matter is whether or not the debtor will succeed in repaying the loan. A smaller risk is when the loan is repaid too early. Our focus was on identifying such loans quarterly. In our thesis, using bank and external macroeconomic data, we constructed several predictive models based on different methods. Our results indicate, considering the constraints we defined, that we managed to identify roughly 50% of such loans.

• Keywords: loan, early repayment, predictive model, classification