An Enhanced Risk Prediction Framework for Blockchain-based Financial Transactions Using Interval Neutrosophic Covering Rough Sets with Heuristic Search

Elvir Akhmetshin^1,2,*, Ilyos Abdullayev³, Erkin Shodiev^4,5, Samariddin Makhmudov^6,7,8, Gavkhar Khidirova⁹, K. Shankar¹⁰

¹Faculty of Economics, RUDN University, Moscow, 117198, Russia

²Khorezm University of Economics, Urgench, 220100, Uzbekistan

³Department of Business and Management, Urgench State University, Urgench, 220100, Uzbekistan

⁴Department of Accounting and Business Management, Mamun University, Khiva, 220900, Uzbekistan

⁵Department of Accounting and Auditing, Tashkent Institute of Irrigation and Agricultural Mechanisation Engineers National Research University, Tashkent, 100000, Uzbekistan

⁶Department of Finance, Alfraganus University, Tashkent, 100000, Uzbekistan

⁷Department of Finance and Tourism, Termez University of Economics and Service, Termez, 190111, Uzbekistan

⁸Center of the Engagement of International Ranking Agencies, Tashkent State University of Economics, Tashkent, 100066, Uzbekistan

⁹Department of Tourism and Hotel Management, Bukhara State University, Bukhara, 200100, Uzbekistan

¹⁰Department of Computer Science and Engineering, Saveetha School of Engineering, Saveetha Institute of Medical and Technical Sciences, Chennai, 602105, India

Emails: elvir@mail.ru; ilyos.a@urdu.uz; shodiyev_erkin@mamunedu.uz; s.maxmudov@afu.uz; g.r.xidirova@buxdu.uz; drkshankar@ieee.org

Abstract

The most efficient device for modelling uncertainty in decision-making issues is the neutrosophic set (NS) and its add-ons, such as NS of complex, interval, and interval complex. An efficient device for establishing uncertainty in decision-making by inserting three grades of truth, indeterminacy, and falsehood of an established statement. Recently, financial globalization has significantly expanded various methods for enhancing service quality using advanced resources. The practical application of the blockchain (BC) model enables stakeholders concerned about the hazard and return prediction models of economic products. To explore the application of deep learning (DL) in processing financial trading data, a neural network (NN) and DL data are utilized. Absolute stock indices and financial data are utilized for analyzing the efficiency of these models in financial prediction and analysis. This paper presents an Enhanced Risk Prediction Framework for Financial Transactions System Using Interval Neutrosophic Covering Rough Sets (ERPFFTS-INCRS) model. The aim is to develop an effective risk prediction model that enhances the reliability and security of BC financial transactions under uncertain conditions, utilizing neutrosophic logic. Initially, the z-score standardization method is used to clean, transform, and organize raw data into a structured and meaningful format. Furthermore, the ERPFFTS-INCRS method implements the INCRS method for the financial classification process. Finally, the hyperparameter selection for the INCRS model is performed by implementing the Elephant Herding Optimisation (EHO) algorithm. The experimental evaluation of the ERPFFTS-INCRS approach is examined under the metaverse financial transactions (MFT) dataset. The comparison analysis of the ERPFFTS-INCRS approach revealed a superior accuracy value of 98.77% compared to existing methods.

Keywords: Risk Prediction Framework; Blockchain; Financial Transactions; Neutrosophic Set; Fuzzy Set; Interval Neutrosophic Covering Rough Sets