Kharazmi University International Journal of Supply and Operations Management 2383-1359 12 3 2025 08 01 Blockchain Applications in Value Added Tax Refund: A Deep Learning-Based Dual-Stage SEM-ANN Analysis 293 317 2959 10.22034/ijsom.2025.110332.3054 EN Thuy-Thanh Thi Tran Vinh University, Vinh City, Vietnam Tri-Quan Dang Ho Chi Minh City University of Foreign Languages-Information Technology 0000-0003-3551-9198 Luan-Thanh Nguyen Ho Chi Minh City University of Foreign Languages-Information Technology 0000-0002-3118-0572 Duc-Viet Thi Dang Posts and Telecommunications Institute of Technology, PTIT, Tran Phu, Ha Dong, Hanoi, Vietnam 0000-0002-4953-4906 Journal Article 2024 01 23 This study examines the determinants of businesses' readiness to implement blockchain technology for value-added tax (VAT) refunds. This study enhances Technology Acceptance Model 2 (TAM2) by incorporating the concept of perceived risk into the framework. This study utilized a two-stage approach that integrated Partial Least Squares Structural Equation Modeling (PLS-SEM) with Artificial Neural Network (ANN) predictive analytics to test the proposed hypotheses. The ANN technique was employed to identify and analyze the potential nonlinear effects within the model. A total of 175 self-administered questionnaires were used for the analysis. The study found that a significant relationship between TAM2s’ constructs and intention to use blockchain technology can markedly enhance the efficiency of VAT refund operations from a managerial standpoint. The theoretical implications of blockchain technology are substantial, as it introduces novel concepts of trust and accountability in tax interactions, disrupts traditional intermediaries, modifies the balance of information, and redefines contract enforcement. http://www.ijsom.com/article_2959_c61cd0b3901493d62133b3e7e8410fb3.pdf