بررسی متغیرهای موثر بر ریسک اعتباری مشتریان حقوقی بانک‌ها با استفاده از ماشین بردار پشتیبان و درخت تصمیم

نوع مقاله : علمی - پژوهشی

نویسندگان

1 استادیار، گروه حسابداری، واحد بابل، دانشگاه آزاد اسلامی، مازندران، ایران.

2 دانشجوی دکتری حسابداری، گروه حسابداری، دانشگاه آزاد اسلامی، مازندران، ایران.

3 استادیار، گروه بانکداری، پژوهشکده پولی و بانکی بانک مرکزی، تهران، ایران.

چکیده

افزایش نسبت مطالبات غیرجاری به تسهیلات اعطایی به‌عنوان شاخص ریسک اعتباری بانک­ها می‌تواند سلامت شبکه بانکی، نظام مالی و اقتصاد حقیقی را به‌خطر اندازد. از این رو در این مقاله، بررسی این ریسک با استفاده از نسبت مانده واقعی مطالبات غیرجاری و با تمرکز بر مجموعه‌ای گسترده از متغیرها شامل متغیرهای مالی، غیرمالی، خصوصیتی قراردادها، حسابرسی و اقتصادی، در نمونه‌ای از 677 پرونده تسهیلاتی مشتریان حقوقی یک بانک دولتی برای سال‌های 1385 تا 1396 مورد توجه قرار گرفت. براساس نتایج حاصله، در انتخاب متغیرهای تاثیرگذار بر ریسک اعتباری، الگوریتم لاسو با عملکرد بهتر به شناسایی 10 متغیر کلیدی از گروه متغیرهای مالی، اقتصادی و حسابرسی منتهی شد. با این وجود نتایج آموزش این ویژگی‌ها توسط مدل ماشین بردار و درخت تصمیم که بیانگر بهترین نتایج در قالب الگوریتم لاسو به همراه درخت تصمیم هستند، ضریب اهمیت اندکی را برای متغیرهای حسابرسی در نظر می‌گیرند. از این رو استفاده از الگوریتم لاسو به همراه درخت تصمیم با تمرکز بر متغیرهای مالی و اقتصادی می‌تواند از کفایت لازم برخوردار باشد.

کلیدواژه‌ها

عنوان مقاله [English]

Investigating the Variables Affecting Banks’ Legal Customers Credit Risk, Using Support Vectors Machine and Decision Tree

نویسندگان [English]

  • Iman Dadashi 1
  • Sajjad Kordmanjiri 2
  • Zahra Khoshnoud 3
  • Hamidreza Gholamnia Roshan 1
1 Assistant Prof, Department of Accounting, Babol branch, Islamic Azad University, Mazandaran, Iran.
2 Ph.D. Candidate in Accounting, Babol branch, Islamic Azad University, Mazandaran, Iran.
3 Associate Prof, Banking Group, Monetary and Banking Research Institute, Tehran, Iran.
چکیده [English]

The increase of non-current debts to lending facilities ratio as an indicator of banks' credit risk can endanger the health of the banking sector, financial system and the real economy. Hence, in this paper, analyzing credit risk through the actual balance of non- performing debts by focusing on a broad set of variables including financial, non-financial, contractual, audit and economic variables in a sample of 677 legal customer facility files of a State Bank for the years 2006- 2017 has been accomplished. Based on the results, the LASSO Algorithm with better performance has identified 10 key financial, economic and audit variables affecting the credit risk. However, training these features by support vector machine and decision tree model, which represent the best results in the Lasso algorithm with the decision tree application, confirms the small significance factor for the audit variables. Therefore, using LASSO algorithm with decision tree and focusing on financial and economic variables can be sufficient for credit risk analysis.

کلیدواژه‌ها [English]

  • Credit Risk
  • Non-Current Debts
  • LASSO Algorithm
  • Decision Tree
  • Recovery Rate

مراجع

1. Addo,P. M., Guegan,D., & Hassani, B., (2018). Credit Risk Analysis Using Machine And Deep Learning Models. Risks, 6(2). 38.
2. Aghaii, M., Rezagholizadeh. M. (2016). Investigating the Factors Affecting the Volume of Overdue and Overdue Receivables of Selected Branches of Sepah Bank. Quarterly Journal of Islamic Finance and Banking Studies. 2(3). 95-111. (in Persian).
3. Bastos, J. A. (2010). Forecasting Bank Loans Loss-Given-Default. Journal of Banking & Finance, 34 (10), 2510-2517.
4. Chow, J. K. (2017). Analysis of Financial Credit Risk Using Machine Learning, Master of Business Administration Dissertation, Aston Business School, Aston University, Birmingham, United Kingdom.
5. Darabi, R. & Mashayekhi, GH. (2016). The Impact of Financial Intelligence on Credit Risk Prediction Using the Support Vector Machine Model. Accounting and Auditing Research. 9 (30). 1-22. (in Persian).
6. Ebrahimi, M. & Daryabar, A. (2011). Credit Risk Management in the Banking System - Data Envelopment Analysis and Logistic and Neural Network Approach. Journal of Investment Knowledge, 1(2):35-62. (in Persian).
7.  Falahpour, S., Raai, R. & Hendijani, M. (2014). Neural Network Approach Based on Artificial Bee Colony for Estimating Credit Rating of Bank Clients. Journal of Financial Engineering and Bond Management.5 (21). 33-53. (in Persian).
8. General Management of Banks and Credit Institutions Supervision. (2005). Principles on Credit Risk Management. Banking Studies and Regulations Office, Central bank of Islamic Republic of Iran. (in Persian).
9. Habibi, R., Kouhi, H. & Baaidi, H. (2018). Bank Facilitation Decisions Using Genetic Algorithm Method (Case Study: Sepah Bank's Real Customers), Islamic Financial and Banking Studies Quarterly. 4 (9). 33-71. (in Persian).
10. Hori, M. & Kaveh, M. (2015). Designing a Model to Predict the Credit Score of Banking Customers Using Multi-criteria Hyper-Algorithm and Ant-Fuzzy-Colony Neural Network Algorithm (Case Study of Tehran Bank Post Branches). Iranian Journal of Management Research. 19 (1). 91-116. (in Persian).
11. Jafari Eskandari. M. & Rouhi, M. (2017). Credit Risk Management of Banking Customers Using Support Vector Machine Optimized by Genetic Algorithm with Data Mining Approach. Journal of Asset Management and Financing. 5(4). 17-32. (in Persian).
12. Kabari, L. G., & Nwachukwu, E. O. (2013). Credit Risk Evaluating System Using Decision Tree – Neuro Based Model. International Journal of Engineering Research & Technology, 2(6), 2738-2745.
13. Karaa, A., Krichene, A. (2012). Credit–Risk Assessment Using Support Vector Machine and Multilayer Neural Network Models: A Comparative Study - Case of Tunisian Bank. Journal of Accounting and Management Information Systems. 11(4), 587-620.
14. Kimura, H.,  Basso, L. &  Kayo, E. (2015). Decision Models in Credit Risk Management. Decision Models in Engineering and Management, Springer, pp.57-73.
15. Mansourfar, Gh., Piri, P., Alikhani, Z. & Asadi, M. (2018). Predicting Financial Distress Given the Moderating Effects of the Independent Auditor's Report. 16th Iranian National Accounting Conference, Isfahan. (in Persian).
16. Mirghafouri, H. & Amin Ashouri, Z. (2015). Credit Risk Assessment of Bank Customers. Two Business Management Quarterly.7 (13). 247-266. (in Persian).
17. Min, J. H.  & Lee, Y-C. (2008). A Practical Approach to Credit Scoring. Expert Systems with Applications, 35, 1762–1770.
18. Mohagheghneia, M., DehghanDehnavi, M. & Baai, M. (2019). The Impact of Internal and External Factors on the Banking Credit Risk in Iranian Banking Industry. Journal of Financial Economics.13 (46). 127-144. (in Persian).
19. Mohammadiyan Haji Kord, A., Asgharzadeh Zaafarani, M. & Emamdoost, M. (2016). Credit Risk Assessment of Corporate Customers Using Support Vector Machine and Genetic Algorithm Hybrid Model - a Case Study of Tejarat Bank. 7 (27). 17-32. (in Persian).
20. Norouzi, P. (2013). The Impact of Large Variables on Credit Risk of Banks in Iran. . Journal of Monetary& Banking Research, 7 (20). 237-257. (in Persian).
21. Office of Economic Research and Policies. (2006-2017). Economic Indicators. Central bank of Islamic Republic of Iran. (in Persian).
22. Office of Banking Studies and Regulations. (2006). Guidelines for Classifying the Assets of Credit Institutions. Central bank of Islamic Republic of Iran. (in Persian).
23. Pérez-Martín, A., & Pérez-Torregrosa, A., & Vaca, M. (2018). Big Data Techniques to Measure Credit Banking Risk in Home Equity Loans. Journal of Business Research, 89, 448-454.
24. Pourkazemi, M., Sedaghatparrast, E. and Dehpanah, R. (2017). Estimation of the Bank's Real Customer Failure Estimation Using Neural Networks Method. (Case Study: Pasargad Bank). Islamic Financial and Banking Studies Quarterly, 6.1-24. (in Persian).
25. Raai, R. & Falahpour, S. (2008) Application of Support Vector Machine in Predicting Corporate Financial Distress by Using Financial Ratios, Accounting and Auditing Reviews. 15 (53). 17-34. (in Persian).
26. Rostami, M., Nabizade, A & Shahi, Z. (2018). Factors Affecting Credit Risk of Commercial Banks of Iran with Emphasis on Banking and Macroeconomic Specific Factors. Journal of Asset Management and Financing.6 (4). 79-92. (in Persian).
27. Rostamzadeh, P., Shahnazi, R. & Neisani, M. (2018). Identification of Factors Affecting on Credit Risk in the Iran Banking Industry of Iran Using Stress Test. Journal of Research in Economic Modeling.8 (32). 91-128. (in Persian).
28. Safari, S., Ebrahimi, M. & Taheri, M. (2011). Credit Risk Management in the Banking System Comparative Approach to Data Envelopment Analysis and Neural Network. Scientific-Research Journal of Shahed University. 9(47). 121-140. (in Persian).
29. Taghipour, M., Saghaiy, A. & Bagheri, M. (2015). Investigating the Factors Affecting the Credit Risk Assessment of Bank Clients Using a Combined Approach to Data Mining Techniques. Conference on Industrial Engineering, Management and Accounting. (in Persian).
30. Tari. F., Ebrahimi, A., Mousavi, J. & Kalantari, M. (2017). Comparison Between Neural Network, Genetic Algorithm and Logit Models in Evaluating Consumer Credit Risk. Journal of Monetary & Banking Research, (34). 657-680. (in Persian).
31. Yao, X. Crook, J, & Andreeva, G. (2015). Support Vector Regression for Loss Given Default Modelling. European Journal of OperationalResearch. 240(2), 528-538.
32. Yanping, Y., Zhengming, Q., Min, Y., Rui, G., Liting, F., Penghui, G. (2012). Research on the Application of Decision Tree to the Analysis of Individual Credit Risk. Information Technology, 25, 209-214.
33. Yu, L., Wang, S. Y., Lai, K. K. (2008). Credit Risk Assessment with a Multistage Neural Network Ensemble Learning Approach. Expert Systems with Applications. 34(2), 1434–1444
34. Wang, G., Ma, J. (2012). A Hybrid Ensemble Approach for Enterprise Credit Risk Assessment Based on Support Vector Machine. Expert Systems with Applications. 39(5), 5325–5331.
 

فایل ها

هم رسانی

ارجاع به این مقاله

آمار