Maximizing Creditworthiness Assessment Fidelity in Incomplete Information Environments

The object of study is a borrower. The subject of study is the borrowers’ creditworthiness. The purpose of the research is to assess and predict borrowers’ creditworthiness in incomplete information environments. The paper proposes an approach to preliminary data processing based on the combination of statistical methods and intelligent techniques, leading to improved accuracy in model prediction and credit risk assessment. During the study, data were collected and analyzed, models for estimating borrowers’ creditworthiness were developed, assessed and compared. The results of the analysis showed increase in quality metrics of models based on data pre-processed with the proposed approach. The practical significance of the study is in the application of the proposed approach in borrower creditworthiness modeling. These models were implemented as part of the advanced IRB-approach in a major Russian bank.