Show simple item record

dc.contributor.advisorPärna, Kalev, juhendaja
dc.contributor.authorÖzay, Tevfik Can
dc.contributor.otherTartu Ülikool. Loodus- ja täppisteaduste valdkondet
dc.contributor.otherTartu Ülikool. Matemaatika ja statistika instituutet
dc.date.accessioned2016-07-08T06:50:50Z
dc.date.available2016-07-08T06:50:50Z
dc.date.issued2016
dc.identifier.urihttp://hdl.handle.net/10062/52455
dc.description.abstractThis study is devoted to small loan evaluation modelling which is known as credit scoring. Credit scoring models help the decision takers (such as credit offices, banks …) decide customers’ creditworthiness in short time without prejudice. Main goal of this master thesis was to understand feasibility and effectiveness of credit scoring model by using logistic regression technique and obtaining important variables for credits scoring models. Furthermore we targeted to reveal how segmentation (creating different score cards for different age groups) can help to predict more accurately. In this study, we worked with real data which was provided by local company in Estonia. In conclusion, our results showed that credit scoring by logistic regression helped to discriminate good customers effectively and the use of segmentation improves the model’s accuracy.en
dc.language.isoenen
dc.subjectregression analysisen
dc.subjectlogistic regressionen
dc.subjectcredit scoringen
dc.subjectkrediidiskooringet
dc.subjectregressioonanalüüset
dc.subjectlogistiline regressioonet
dc.subject.othermagistritöödet
dc.titleCredit scoring by segmented modellingen
dc.typeThesisen


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record