Primena logističke regresije u proceni kreditnog rizika / Application of logistic regression in credit risk assessment
2019
Tekić, Dragana (https://orcid.org/0000-0002-1924-6196) | Mutavdžić, Beba (https://orcid.org/0000-0002-7631-0465) | Novković, Nebojša (https://orcid.org/0000-0003-2419-5765) | Milić, Dragan (https://orcid.org/0000-0002-4283-4068) | Zekić, Vladislav (https://orcid.org/0000-0002-7377-2402) | Novaković, Tihomir (https://orcid.org/0000-0002-8405-3403)
The success of a financial institution's business depends on its ability to predict and quantify risk, and to assess as closely as possible the creditworthiness of each potential client. In the work that follows, a logistic regression method is applied to determine what factors, which financial analysts consider when interpreting a loan request, influence the decision to approve a loan. Specifically, the factors that are assumed to have an impact on credit risk, to make a decision on accepting or rejecting a loan application are: current account lockout, last year account lockout, total indebtedness rate, bank indebtedness rate, total liquidity ratio, credit exposure, return on equity and return on assets. The results of the research show that the most important factors included in the final model are the account lockout, the account lockout in the last year and credit exposure, and credit exposure is the most important predictor. The model successfully classifies 86% of all cases.
显示更多 [+] 显示较少 [-]