Korištenje regresijskih metoda u kreditnom skoringu

master's thesis

• Author: Jana Rajčić
• Mentor: Siniša Slijepčević (mentor)
• Committee member: Siniša Slijepčević (predsjednik povjerenstva)
• Committee member: Miljenko Huzak (član povjerenstva)
• Committee member: Vanja Wagner (član povjerenstva)
• Committee member: Bojan Basrak (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2019-09-23, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

Cilj ovog rada je bio upoznati se sa linearnom regresijom, te kako korištenjem regresijskih modela možemo odrediti klijentovu vjerojatnost otplaćivanja obveze na temelju prijašnjih podnositelja zahtjeva. Glavna pitanja su bila koliko bi trebao biti veliki uzorak s kojim radimo i kako uopće treba izgledati podjela na „dobre“ i „loše“ klijente, odnosno kako glasi definicija „dobrih“ i „loših“ klijenata, te je bilo potrebno odrediti razumnu povijest otplate. U radu smo spomenule neke od statističkih metoda za izgradnju kreditne bodovne kartice. Zaključak rada je da bez obzira koju definiciju „dobrih“, odnosno „loših“ odaberemo, ne postoji učinak na metodologiju kreditne bodovne kartice jer se ona zapravo razvija na temelju definicija.

Abstract (english)

The aim of this work was to introduce the notion of linear regression, and explain how one can, by using regression models, determine the probability of a customer to pay off an obligation, based on behavior of previous applicants. The main questions were on how large the sample for analysis should be, and how we should divide data into "good" and "bad" clients, i.e. what does the definition of "good" and "bad" clients looks like, and it was necessary to determine a reasonable history of repayment. In the thesis we mentioned some of the statistical methods for building credit score cards. The conclusion of the work is that no matter what definition of "good" or "bad" we choose, there is no effect on the methodology of credit score cards because it actually develops on the basis of definitions.

• Keywords: linearna regresija; regresijski modeli; izgradnja kreditne bodovne kartice
• Keywords (english): linear regression; regression models; building credit score cards
• Language: croatian
• URN:NBN: urn:nbn:hr:217:964172
• Study programme: Title: Finance and Business Mathematics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-02-18 12:46:06