Construction and application of scoring models

This paper focuses on the definition and practical application of scoring models that are used not only in the bank sector, but also by other institutions to assess whether the client can get a loan or not. If there is an incorrect decision due to wrong definition in this model then the institution is exposed to risk of loss. Therefore when any application for the loan is posted, institution needs to do the best quantification of the risks. Only after that is able to decide, on the basis of available information, about the applicant. There will be used logistic regression method in the process of constructing the models in this paper. The output of this method has to estimate the conditional probability of client repayment of the loan with given characteristics. One of the most studied characters of the scoring model is the ability of diversification represented by the Lorenz curve and quantified by Gini coefficient.