Differential-Evolution–Tuned XGBoost for cyber financial fraud detection in Omani SMEs: An empirical evaluation under Vision 2040

Read full text
Authors: Raghad Yaqoob Al-Julandani
Cited by: 0

Abstract

In the past decade, small and medium enterprises (SMEs) in Oman have witnessed significant change, but this change has led to increased exposure to electronic financial fraud. Recent reports emphasize the importance and need for adaptive artificial intelligence (AI) and cybersecurity frameworks that are aligned with Oman Vision 2040. The research is based on the development of the Differential Evolution Model – Controlled Extreme Gradient Enhancement (XGBoost) as an improved AI method for SMEs to identify electronic fraud transactions since current systems are based on limited rules in detecting emerging attack patterns to reduce false positive alerts. The model uses differential evolution (DE) with a metaheuristic optimization technique to help tune the XGBoost hyperparameters. To replicate real-life mobile money transactions and credit card fraud to measure the data set, the study uses the PaySim emulator. During the study, a blockchain-inspired audit component was explored to enhance trust and transaction traceability. The XGBoost baseline was compared to the initial results, showing a 20% reduction in false positives and a 15–18% improvement in accurate recall (AUC-PR). Due to confidentiality restrictions, real verification of SME transactions was not applied in this study and therefore the research is limited due to its reliance on simulation data only. In future studies, to maintain privacy and enhance generalizability, deep learning models, standardized learning frameworks, and the integration of real-world financial data sets will be combined. The proposed model for detecting financial fraud in SMEs promotes the intersection of modern technologies (artificial intelligence, blockchain, and cybersecurity) for flexible, innovation-driven digital finance, in line with Oman Vision 2040. Keywords: Artificial Intelligence, Differential Evolution, XGBoost, Cyber Financial Fraud, Blockchain, Cybersecurity, FinTech, SMEs, Imbalanced Learning, Oman Vision 2040.

Fields of Study & Subjects

Imbalanced Data Classification Techniques Privacy-Preserving Technologies in Data Network Security and Intrusion Detection

Publication details

Journal / source
International Journal of Management & Entrepreneurship Research
Volume
8
Issue
2
Pages
249–262
Publication date
16 February 2026
Language
EN