تعیین ماندگاری و سودآوری مشتریان بانک با استفاده از تعمیم مدل درخت تصمیم و جنگل رگرسیون

نوع مقاله : مقاله پژوهشی

نویسندگان

1 دانشیار، گروه مدیریت، دانشکده مدیریت و حسابداری، دانشگاه علامه طباطبایی، تهران، ایران

2 استادیار، گروه بانکداری، دانشکده بانکداری، موسسه عالی آموزش بانکداری ایران، تهران، ایران

3 کارشناسی ارشد بانکداری، دانشکده بانکداری، موسسه عالی آموزش بانکداری ایران، تهران، ایران

چکیده

در این تحقیق با استفاده از روش‌های جنگل تصادفی و جنگل رگرسیون که تعمیم‌یافته مدل درخت تصمیم و رگرسیون هستند، عوامل مؤثر بر ماندگاری و سودآوری مشتریان ارزی یک بانک تجاری دولتی، موردبررسی قرار می­گیرد. جامعه آماری تحقیق شامل مشتریانی است که دارای یکی از انواع حساب‌های ارزی بانک بوده و به‌طور همزمان از خدمات حواله‌های ارزی، گشایش اعتبارات اسنادی و تسهیلات ارزی استفاده می­نمایند. همچنین صحت نتایج با رگرسیون لجستیک و بگینگ مقایسه می­شود. در روش جنگل رگرسیون، با استفاده از معیار درصد میانگین قدرمطلق خطا (MAPE) ، به مقایسه دقت نتایج مدل با روش رگرسیون خطی پرداخته می­شود. سپس درجه اهمیت هر یک از متغیرهای مستقل بر روی متغیرهای وابسته خرید بعدی، کاهش سطح فعالیت، کاهش سودآوری و تداوم سودآوری تعیین می‌شود.
نتایج نشان می­دهد در صورت ارائه تسهیلات ارزی بیشتر به مشتریان تولیدی و درنتیجه، گشایش اعتبارات اسنادی یا انجام حواله‌های بیشتر برای این قبیل از مشتریان، احتمال افزایش دوره ماندگاری مشتریان افزایش می­یابد. همچنین افزایش مبلغ تسهیلات ارزی ارائه‌شده، حواله­های انجام‌شده و همچنین اعتبارات اسنادی گشایش شده به‌واسطه نرخ سود بازپرداخت تسهیلات، کارمزدها و سود حاصله از فروش ارز و ضمانت‌نامه‌ها در بالا بردن میزان سودآوری مشتریان نقش بسزایی دارند.

کلیدواژه‌ها


عنوان مقاله [English]

Determining Retention and Profitability of Bank Customers Using Extended Decision Tree and Forest Regression

نویسندگان [English]

  • Mohammad Taghi Taghavifard 1
  • Reza Habibi 2
  • Mojtaba Aghaei 3
1 Associate Professor of Industrial Management, Faculty of management and accounting, Allameh Tabataba'i, Tehran, Iran
2 Assistant Professor, Faculty of banking, Iran Banking Institute, Tehran, Iran
3 Master of Science, Faculty of banking, Iran Banking Institute, Tehran, Iran
چکیده [English]

In this paper, effective factors on retention and profitability of customers a state owned bank was studied using random forests and regression forests methods. The data was collected form Bank Sepah database. The statistical population consists of 169 corporations holding different types of deposit accounts and utilized one of the different types of services such as sending and receiving payment orders, letters of credit (LC) and foreign exchange facilities during the research period (1389-1391), simultaneously.
In this paper, the accuracy of the results of random forests method is compared with the results of logistic regression and bagging methods using area under the Receiver Operating Characteristic (ROC) curve. Furthermore, the results of regression forests method was compared to those of linear regression by calculating Mean Absolute Percentage Error (MAPE). Then, we tried to determine the importance of effective independent variables on dependent ones, i.e. next purchase, activity defection, profit drop and profit continuity using random forests and regression forests.
The results show that in the case of offering currency facilities to the customers active in the production fields leading to opening LCs and receiving more payment orders will increase the probability of customer retention. In addition, an increase in the amount of foreign exchange facilities and payment orders offered by banks due to the rate of return of foreign exchange facilities, banking fees, income resulting from issued warranties and selling currencies plays an important role in the profitability of customers.

کلیدواژه‌ها [English]

  • Customer Retention
  • Profitability
  • Bank
  • Decision tree
  • Regression Forest
[1] Meltzer, M. (2007). A CRM approach to customer profitability. Downloadable www.amt.eu.com.
[2] Gupta. S. (2007). A model to determine customer lifetime value in a retail banking context. European Management Journal 25: 221-234.
[3] Dennis, D. (1998). Customer loyalty strategies. Journal Consumer Marketing 5: 434-448.
[4] Hampe J., and Swatman, F. (2002). Customer relationship management. Case studies of  five Swedish. Technical report. Luella University of Technologies. Sweden.
[5] Stahl, F., Heitmann, M., Lehmann, D. R., and Neslin, S.A. (2012). The impact of brand equity on customer acquisition, retention, and profit margin. Journal of Marketing 11: 44-63.
[6] Siu N. Y. M., Zhang, T. J. F., and Yau, C. Y. J. (2013). The roles of justice and customer satisfaction in customer retention: a lesson from service recovery.  Journal of Business Ethics 114: 675-686.
[7] Park, S. H., and Han, S. P. (2014). From accuracy to diversity in product recommendations: relationship between diversity and customer retention. International Journal of Electronic Commerce 18:51-71.
[8] Ganesh, J., Arnold, M. J., and Reynolds, K. E. (2000). Understanding the customer base of service providers: An examination of the differences between switchers and stairs. Journal of Marketing 64: 65–87.
[9] Jain R., Jain S., Dhar U. (2003). Measuring customer relationship management. Journal of Service Research 2: 97 -109.
[10] Gilbert, C. D., and Karen, C. C. (2003). Relationship marketing practice in relation to different bank ownership: A study of banks in Hong Kong. Technical report. Department of Economics. National University of Hong Kong. Hong Kong.
[11] Duda, R.O. Hart, P.E. and Stork, D.G. (2001). Pattern classification. Wiley.
[12] Breiman, L. and Adele, C. (2013). Breiman and Cutler’s random forests for classification and  regression. MASS 4: 6-17.
[13] Payet, N., and Todorovic, S.(2008). Random forest random field. Technical report. School of Electrical Engineering and Computer Science Oregon State University. USA.
[14] Clemencon, M. Depecker, and N. Vayatis. (2013). Ranking forests. Journal of Machine Learning Research 14: 39–73.
[15] Denil, M., Matheson, D. and de Freitas, N. (2013). Consistency of online random forests. In Proceedings of the ICML Conference. arXiv:1302.4853.
[16] Scornet. E. (2014). On the asymptotic of random forests. arXiv:1409.2090.
[17] Scornet, E., Biau, G. and Vert, J. P. (2015).  Supplementary materials for : Consistency of random forests. arXiv:1510.2180.
[18] Biau, L., Devroye, M. and Lugosi. G. (2015). Consistency of random forests and other averaging classifiers. Journal of Machine Learning Research 9: 2025–2033.
[19] Abrazi, M., Abbasi, A., HaghShenas, A., and Ansari, M. E. (2009). The effect of the proportion of individuals, organizations and organizational culture on organizational behavior faculty members (Case Study Shaid Bahonar University). Management Researches in Iran 14, 1-20. (In Persian).
[20] Sepehri, M., Noorouzi, A., Teymourpour, B., and Choobdar, S. (2012). Discovering the reasons of customer defection of e-banking services. Data mining methods. Management Researches in Iran 17: 139-160. (In Persian).
[21] Rezapour, M., Sepehri, M. and Rezapour, H. (2014). Determining the best methods of assessing learners in e-learning courses. Management Researches in Iran 20:145-161. (In Persian).
[22] Sohrabi, B., Reisi, I., and Zare, F. (2017). Designing systems to optimize the management of bank lending based on clustering and classification methods. Modern Researches in Decision Making 2:53-76. (In Persian).
[23] Karimi, M. and Kazemi, M. A. (2017). Predicting failure and plan maintenance of ATM. Data mining method. Modern Researches in Decision Making 3: 113-129. (In Persian).
[24] Hanley, J. A., McNeil, B. J. (1982). The meaning and use of the area under a receiver operating  characteristic (ROC) curve. Radiology 143: 29–36.
[25] Efron, B and Tibshirani. H. (1993). An introduction to the bootstrap. Chapman and Hall.
[27] Soderlund, M., and Vilgon, M. (1999). Customer satisfaction and link to customer profitability. SSE/EFI Working Paper Series in Business Administration No.1999:1. USA.
[28] East, R. (2005). Consumer loyalty: singular, additive or interactive? Australasian Marketing Journal 13: 22-35.
[29] Lindgreen A., Palmer R., Vanhamme J., Wouters J. (2006). A relationship management assessment tool: Questioning, identifying, and prioritizing critical aspects of customer relationships. Industrial Marketing Management 35: 57-63.
[30] Brewton J. and Schiemann W. (2003). Measurement: The missing ingredient in today's CRM strategy. Cost Management 17: 5 -14.
[31] Vens, C. and Costa, F. (2011). Random forest based feature induction. Technical report. Department of Computer Science. Katholieke Universiteit. Belgium.