نوع مقاله : مقاله علمی پژوهشی
نویسندگان
1 علوم اقتصادی ،مالی-علوم بانکی ، واحد علوم تحقیقات ، دانشگاه بین الملل ، دانشگاه آزاد- واحد قشم ، تهران ، ایران
2 استادیار، مدیریت مالی، واحد قشم، دانشگاه آزاد اسلامی، قشم، ایران .
3 دانشیار، حسابداری و مدیریت مالی، دانشکده علوم انسانی، واحد بندر عباس، دانشگاه آزاد اسلامی، بندر عباس، ایران
4 دانشیار، مدیریت بازرگانی – مدیریت مالی، واحد اسلامشهر، دانشگاه آزاد اسلامی، تهران، ایران
چکیده
کلیدواژه‌ها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Abstract
Objective:
To score customers effectively, it is essential to establish a fair and appropriate scoring system. This system should classify customers into different categories based on credit and behavioral criteria and assign them scores aligned with their performance. Moreover, developing methods to evaluate and monitor customers over time is necessary. The objective of this study is to Identification and analysis of credit and behavioral indicators and propose a model for ranking customers in unsecured small loans. Customer credit evaluation is a complex process that involves reviewing documents, analyzing financial status, and assessing the customer's payment history. A critical aspect of this evaluation is determining the customers' ability, willingness, and capacity to repay loans. This research aims to present a credit evaluation and ranking model for unsecured small loans and estimate the probability of default in the bank’s digital banking system.
Methodology:
The statistical population for this research includes all retail customers of the digital banking services of Khavar-e-Miyaneh Bank. A census method was used to study the customers during the fiscal years 1400–1401. The model developed by Ahmadi Koosha and colleagues (1403) was used as the basis for ranking customers in unsecured small loans. Indicators such as age, loan amount, eligible loan request amount, total loans received, score, anti-money laundering approval, gender, occupation, city, education level, loan status, and job type were identified as initial input variables. Data analysis was conducted using fuzzy regression and inspired by Python libraries like SciKit-Fuzzy and NumPy, resulting in a fuzzy regression formula. To finalize the evaluation of the credit scoring and ranking model, an example with a customer was presented.
Findings:
The results indicate that variables such as the requested loan amount, total loans received, and occupation do not influence the repayment of installments 60 days past due. The ranking of significant variables affecting repayment within 60 days after the due date is as follows: loan settlement status, anti-money laundering status, gender, education level, age, score, job type, and city. Additionally, the model’s accuracy and predictive power were tested using a hypothetical customer, yielding a score of 0.8738. This score demonstrates that the customer is in a favorable position for repayment within 60 days past due and highlights the model's efficiency in credit risk assessment and management, which is crucial for financial institutions to ensure sustainable lending practices.
Conclusion:
Given the identified variables influencing post-due payment behavior, it is vital for financial institutions to incorporate these factors into their risk assessment frameworks. The bank should prioritize implementing stringent anti-money laundering protocols and regularly update its risk assessment criteria to align with regulatory requirements. Furthermore, to maximize the model’s applicability, Khavar-e-Miyaneh Bank should invest in continuous monitoring and refinement of the model, leveraging the latest customer behavior data to adapt to market dynamics effectively.
کلیدواژه‌ها [English]
)