•  
  •  
 

Abstract

Background: Fraud is a pervasive worldwide problem that is evolving rapidly along with the technological advances and causing significant financial losses in a variety of industries. Traditional detection techniques often fall short in the face of more complex and digitalized fraud schemes. The pressing need for more intelligent detection systems is what motivated this review, which attempts to systematically assess the body of research on fraud detection and pinpoint dominant fraud sectors, types, and methods. Methods: This study employed a Systematic Literature Review (SLR) approach to comprehensively assess the current landscape of fraud detection research across multiple domains. Guided by the PRISMA framework, a structured search was conducted using databases such as IEEE, ACM, Springer, and others available through the Saudi Digital Library. Inclusion criteria targeted English-language, peer-reviewed articles published between 2015 and 2024 that focused on fraud detection or prevention. From an initial pool of 848 studies, 75 met all eligibility requirements and were included in the final analysis. Key information such as fraud type, sector, detection methodology, and publication trends was extracted and synthesized to answer four predefined research questions. Result: The analysis showed that most studies focused on financial fraud, with increasing research on e-commerce and healthcare. Credit card fraud and cyber scams were the most frequent topics. Machine learning dominated the methods used, with growing interest in deep learning models like GNNs and transformers. Most studies targeted fraud detection, while prevention received less attention. Common challenges included data imbalance and limited cross-domain applicability. Conclusion: The review highlights a fragmented but rapidly evolving research landscape in fraud detection. While significant technical progress has been made, especially in AI-based detection, most existing models are tailored to specific fraud types or datasets, limiting their applicability across domains. The findings point to a need for more adaptable, cross-sector models and greater emphasis on fraud prevention strategies, which are currently underexplored. Furthermore, the review identifies a lack of standard evaluation practices, making it difficult to compare performance across studies. Addressing these issues will require interdisciplinary collaboration, better data sharing, and an expanded focus on real-world deployment to ensure fraud detection research translates into practical, scalable solutions.

Keywords

Fraud detection, Fraud prevention, Systematic literature review, Machine learning, Deep learning, Finance, E-comm, Healthcare

Article Type

Article

First Page

40

Last Page

66

Publication Date

6-30-2026

Download

Share

COinS