Graph‑Eenhanced fraud detection in health insurance: integrating centrality metrics and community analysis for improved accuracy

Document Type : Full Length Article

Authors

1 Master's Degree, Department of Engineering Sciences, University of Tehran, Tehran, Iran

2 Assistant Professor, Department of Modern Insurance Technologies, Insurance Research Center, Tehran, Iran

3 Assistant Professor, Department of Engineering Sciences, University of Tehran, Tehran, Iran

Abstract

The dynamic nature of fraud and the emergence of new fraudulent methods in the insurance industry, along with the insufficient capability of insurance companies to prevent, identify, and combat it, may lead insurance companies toward bankruptcy in the not-too-distant future. Given that insurance companies serve as the main entity providing insurance services and compensating losses, their significant role in preventing and detecting fraud makes the various strategies they employ to counter fraud significant. In the current situation, identifying and analyzing the phenomenon of fraud, which leads to increased costs in the insurance industry, appears to be essential for controlling factors and ensuring the survival of insurance companies. This article examines the use of graph theory for fraud detection in the health insurance industry. By extracting information from relevant databases and constructing a comprehensive network graph, existing patterns are analyzed and suspicious fraudulent cases are identified. The proposed method was implemented on real data, and the results showed that this method is capable of detecting fraud with 95% accuracy, which is an improvement over the existing method. All implementation codes and supplementary materials are openly available on GitHub [https://github.com/MohammadMehdi41/fraudDetection.git]

Graphical Abstract

Graph‑Eenhanced fraud detection in health insurance: integrating centrality metrics and community analysis for improved accuracy

Keywords

Main Subjects

Journal of Discrete Mathematics and Its Applications
Volume 11, Issue 2
June 2026
Pages 131-151

Files

History

  • Receive Date: 11 December 2025
  • Accept Date: 16 March 2026
  • First Publish Date: 01 June 2026
  • Publish Date: 01 June 2026

Share

How to cite

Statistics