A Predictive Analytics for Customer Lifetime Value Estimation in Digital Banking using Interval-Valued Neutrosophic Set with Fine Tuning Approach

Alisher Sherov^1,2, Ziyodulla Khakimov³, Yurii Vorobev^4,*, Emil Hajiyev⁵, Tatyana Khorolskaya⁶

¹Department of Economics, Mamun University, Khiva, 220900, Uzbekistan

²Department of Finance and Tourism, Termez University of Economics and Service, Termez, 190111, Uzbekistan

³Department of Management and Marketing, Alfraganus University, Tashkent, 100000, Uzbekistan

⁴Department of Economics and Finance, Financial University under the Government of the Russian Federation, Moscow, 125167, Russian Federation

⁵Department of Business Management, Azerbaijan State University of Economics (UNEC), Baku, AZ1001, Republic of Azerbaijan

⁶Department of Money Circulation and Credit, Kuban State Agrarian University named after I.T. Trubilin, Krasnodar, 350044, Russian Federation

Emails: sherov_alisher@mamunedu.uz; z.xakimov@afu.uz; ynvorobev@fa.ru; hajiyev.emil@unec.edu.az; xorolskaya.t@edu.kubsau.ru

Abstract

As a generality of fuzzy sets (FS) and intuitionistic FS (IFS), neutrosophic sets (NS) was progressed by F. Smarandache for signifying incomplete, inaccurate, and uneven data present in the real world. Neutrosophic Logic (NL) is a neonate research field in which every proposition was projected to have the proportion of truth in a sub-set T, I, and F. Neutrosophic sets (NS) have been well employed for indeterminate information handling, and determine benefits to tackle indeterminate data. A NS is categorized by indeterminacy-, truth-, and a falsity- membership functions. Atanassov as a major simplification of FS presented the notion of IFS. IFS are very beneficial in conditions when problem description by linguistic variables, assumed with only a membership function, appears to be difficult. In recent times, IFS have been employed to numerous areas like medical diagnosis, logic programming, decision-making issues, etc. An interval NS (INS) is an example of NS, which is employed in real engineering and scientific applications. Owing to the competition in the banking industry and the importance, access to customer information is vital to establish a successful relationship that benefits both parties. Representing longer-term customer relationships and building brand equity are essential in modern banking, and therefore increasing relationship quality plays a significant part in the development of new services and customer lifetime value (CLV) approximation.  CLV is an estimated profit that can be achieved by the organization from a customer for some time. Presently, the development of Machine Learning (ML) methods has resulted in better precision and effectiveness. Therefore, by utilizing ML methods of real-time customer data, predictions of a more precise future value of the customer are gained by businesses, which helps in establishing a more personal marketing approach. In this manuscript, we propose a Customer Lifetime Value Estimation using Interval-Valued Neutrosophic Set and Parameter Optimization Algorithms (CLVE-IVNSPOA). The foremost main of this paper is to progress a predictive analytics model for estimating customer lifetime value in digital banking utilizing advanced optimization methods. Initially, the data pre-processing phase was employed by using the Z-score method. Moreover, the pelican optimization algorithm (POA) is mainly executed by the feature subset selection in order to select the most optimal features from a dataset. For CLV prediction, the Interval-Valued Neutrosophic Set (IVNS) technique is exploited. At last, the model parameter adjustment process is performed through improved shark optimization (ISHO) algorithm for improving the prediction performance. The experimental evaluation of the CLVE-IVNSPOA occurs using benchmark database. The experimental outcomes indicated out an improved performance of CLVE-IVNSPOA compared to existing systems.

Keywords: Customer Lifetime Value; Neutrosophic set; Digital Banking; Interval-Valued Neutrosophic Set; Improved Shark Optimization Algorithm; Neutrosophic Logic