Original scientific paper

Application of decision trees in credit scoring

Ljiljanka Kvesić

Abstract

Banks are particularly exposed to credit risk due to the nature of their operations. Inadequate assessment of
the borrower directly causes losses. The financial crisis the global economy is still going through has clearly
shown what kind of problems can arise from an inadequate credit policy. Thus, the primary task of bank
managers is to minimise credit risk. Credit scoring models were developed to support managers in assessing the creditworthiness of borrowers. This paper presents the decision tree based on exhaustive CHAID
algorithm as one such model. Since the application of credit scoring models has not been adequately explored in the Croatian banking theory and practice, this paper aims not only to determine the characteristics
that are crucial for predicting default, but also to highlight the importance of a quantitative approach in
assessing the creditworthiness of borrowers.

Keywords

credit scoring; model; decision tree; exhaustive CHAID algorithm

Hrčak ID:

116135

URI

https://hrcak.srce.hr/116135

Publication date:

31.12.2013.

Article data in other languages: croatian

Visits: 2.671 *