|
Modeling Credit Thresholds in Banks: A New Approach Based on the BN Artificial Intelligence Algorithm
|
Mehrdad Jeyhoonipour *1   , Somayeh Azami1 , Sohrab Delangizan1 |
| 1- University of Razi |
|
|
Abstract: (165 Views) |
In the complex regulatory and competitive environment of the banking industry, continuous improvement of internal rating models has become a strategic necessity for financial institutions. Despite significant advances in statistical methods and machine learning algorithms, optimal allocation of credit and determination of credit limits for applicants remain fundamental challenges for risk and credit managers. This paper presents a novel statistical approach based on artificial intelligence that allows estimating the maximum allowable credit and identifying the factors affecting it. The purpose of the research is to determine the threshold limits for granting credit, based on the bank's credit policies, by determining the acceptable range of the probability of default and the amount of risk it accepts in granting credit. In this study, two scenarios with different causal relationships between variables affecting default are examined, which helps bank managers in making informed decisions. The presented model is a tool for calculating the probability of default and determining the credit that can be allocated. The output of this model is designed to be easily implemented in bank branches and to help managers optimize credit decisions.
|
Article number: 4 |
| |
|
|
|
Type of Study: Empirical Study |
Subject:
Monetary Policy, Central Banking, and the Supply of Money and Credit (E5)
Received: 2025/03/15 | Accepted: 2025/05/25 | Published: 2025/06/5
|
|
|
|
|
|
|
| Send email to the article author |
|
|
